Zen Performance Demo vs Intel’s $1000 Broadwell-E i7 6900X

It's over, intel is finished.

>>56149579
Since it's all an internal AMD demo, take it with a grain of salt, but I believe it. Intel is stagnant and AMD had a legendary engineer for ZEN and is now under good management. Two node jumps delivering that much more performance isn't exactly hard to believe.

I'm still waiting for independent reviews, but even if it is a little less than what they have in this benchmark, I'm sold. The chips will be cheaper and perform well, and the boards will do exactly the same fucking shit as Intel X boards and be half the fucking price. I'm more hyped for that to be honest.

>>56149949
this

I'm a massive amd fan, but only an idiot would take a press release at face value.

https://twitter.com/FPiednoel/status/766334621550161920
https://twitter.com/FPiednoel/status/766346666433576961

AYYMD is cheating in Blender & got busted

>tfw I haven't replaced my 2500k yet
I can only wish the best of luck to AMD, I'm tired of seeing all these overpriced cpu for only 5-10% difference.

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

71KB, 552x661px

>>56150018
DELETE THIS RIGHT NOW

>>56149949
>Two node jumps delivering that much more performance isn't exactly hard to believe.
So much this.

I'm going to wait until after the release and further benchmarks are done to see if I want to buy a Zen CPU.

But honestly I would be happy if AMD could match Intel's mid range for a better price. Would really start some shit up.

How much? $2000?

File: 098345098.png (496KB, 738x590px) Image search: [Google]

496KB, 738x590px

>>56149579
https://twitter.com/AMD/status/766408132860010496?lang=en

>>56150146
You'd see Intel pricedrops, but they're at the point where RnD is getting absurd, even for them. AMD will catch up eventually just by the nature of the physical limits of silicon.

I'll monitor it at the very least. Fuck what some twitter fag says, and fuck press releases. Reserving judgement for real world benchmarks and reviews.

>>56150018
Do you know how to code? There's literally nothing wrong with it. Is he really sperging about the name of the variable?

>>56150018
Now, I'm not that advanced at programming or anything, but isn't it basically just checking the vendor ID and cache size?

What's the problem here? I don't see any code that has it do anything significantly different depending on vendor like Intel's compilers used to do.

>>56150054
beat my e8400 famalam.

>>56150018
>implying this sort of practice is uncommon.

>>56150367

His argument is that it basically gimps the Intel chip by recompiling for Zen, making it more focused and efficient for AMD then Intel. He's sperging out because he doesn't understand that its not an Intel architecture vs AMD architecture comparison, it's an IPC comparison. Emphasizing that they have not only matched Intel but its close enlugh that they may have surpassed them.

File: 1470535174297.png (413KB, 736x665px) Image search: [Google]

413KB, 736x665px

>>56149579
I really hope that isn't their highest end cpu

>>56150018
Blender is open source I'm pretty sure they weren't really hiding anything from the beginning and someone was probably making a joke

>>56149579
It's not performance demo, but an IPC demo. Big difference, anon.

>it's legit? It's not a scam like the rx480 cf?
proof of it being scam? Official benchmarks are always done in a safe way, they just cherry pick them really really well.
They are not stupid, they know a real lie could cost them in court.

>>56150978
It isn't this is just a demo of IPC. Rumor mill has high end at either 6 or 8 cores, possibly higher.

>>56150018
You guys are like fucking drones, you don't even know what you're reading.

>>56149579
>A chip which isnt going to come out this year outperforms a chip which already is
>in tests which are unable to be reproduced by anyone else

>>56149579
gib benchmarkds

>>56149579
>Trusting the same website that claimed the 480 will compete with the 480, when it can barely compete with the cut down 1060.

Yeah, it'll probably perform about as well as a sandy lake i3

>>56149579
Marketing tests are always cherry picked

but at the very least zen should be decent competitive

>>56151390
I'm sure it'll be a whole 5-10% faster than Bulldozer

>>56150018
how dare they query the cache size

File: 1456679793007.jpg (181KB, 1280x720px) Image search: [Google]

181KB, 1280x720px

>they use Blender as a demo
>they took YEARS to fix their buggy OpenCL compiler that made using Cycles on an AMD GPU impossible
I'm mad.

Still pretty hyped for Zen.

Does this mean I can use amd unironically and non budget

File: 1460578789623.png (536KB, 757x615px) Image search: [Google]

536KB, 757x615px

>>56150054
>somehow my 5930K is only 5-to-10% faster than your 2500K

File: 900x900px-LL-a2d2ea08_530f0d335d.png (28KB, 727x519px) Image search: [Google]

28KB, 727x519px

>>56151647
If you're a cuck. AMD processors always had shit memory controllers. I doubt zen can even run 2666 ddr4.

>>56151733
What GPU was that even running with?

>>56151733
Well looks like I'll be the average consumer and buy what's better, got nvidia is such a fuck company but their products are just better by most standards, same with Intel

>>56150018
>>56150127
retard

>>56151782
keepk cryin bithc nigga

>>56151695
Garbage movie

>>56151733
>always had shit controllers
And AMD pioneered multi level chip caching
They have a new chip interconnect which is state of the art

>>56151763
You say that example, but RX480 prices are inflated due to high demand and still selling at nearly the same rate as 1060

Anyways I'm more excited to see after Zen and Vega, are those monster HPC APUs a reality or not?
HBM2 on mobile chips in Q4'17/Q1'18 would be interesting but add 3-5w to the package alone, plus we have seen P11 and know how it would perform.
etc.etc.

>>56151841
>RX480 prices are inflated due to high demand

NEIN

>>56151841
>RX480 prices are inflated due to buttcoin miners
ftfy

File: FB_IMG_1471549432756.jpg (47KB, 640x640px) Image search: [Google]

47KB, 640x640px

>>56150018
Wew lad.

>>56149579
I hope its true.
would be interesting seeing what Intel does if AMD actually is successful.

>>56151858
It's inflated because amd is fucking retarded.

>hey guys let's make an inferior product and price it the same because we need to remove our poorfag stigma
>yes this plan will indubitably defecate on the competition

>>56149579
>Intel releases overpriced Broadwell-E which barely outperforms Skylake in certain non-gaming applications
>AyyMD btfos Broadwell-E
it's over anons, Intel is done

>>56151733
>AMD processors always had shit memory controllers
lol retard

>>56151903
ONLY ON AMAZON/EBAY AND OTHER MEME WEBSITE

>>56151923
Name one retail amd cpu that supports ddr4

>>56151903
AMD did nothing wrong with the 480

>>56151939
Summit Ridge

>>56150192
2.1ghz 1060trash vs 1.2mhz 480

480 destroys 1060 in dx12 and dx11 doesnt matter because fps is high enough for everything.

nvidia ADD MOAR MHZ SO WE STAY IN THE GAME LELELELEKEKEK

So what's stopping me from buying 4 titans

>>56151980

>average AMD fanboy

>>56151980
...with a lower power draw...

>>56152068
Because people care about a few watts? Are you retarded?

>>56152119
>amdoomed

http://promotions.newegg.com/msi/16-5275/index.html?cm_sp=Homepage-Top2016-_-P2_msi%2f16-5275-_-http%3a%2f%2fpromotions.newegg.com%2fmsi%2f16-5275%2f1920x360.jpg&icid=366393

Why the hell you need 14nm if it's still a housefire...

File: amd.jpg (138KB, 653x726px) Image search: [Google]

138KB, 653x726px

>>56152155
DEL

File: poo.png (38KB, 1971x697px) Image search: [Google]

38KB, 1971x697px

>>56152155
BTFO

>>56152155
>$5000
no thanks

File: 1471318625325.png (47KB, 1971x697px) Image search: [Google]

47KB, 1971x697px

>>56152215

File: 1471481562966.png (51KB, 702x439px) Image search: [Google]

51KB, 702x439px

>>56152119
People like slim gayming laptops

https://www.youtube.com/watch?time_continue=35&v=QiwfZApg9JY

Demo vs a downclocked Intel cpu.

Even if it really gets the entire 40 percent gain, that still won't let it catch current intels in single core, let alone next year's Intel models.

>>56152239
Post your battlestation

>>56152278
it doesn't matter, if amd can offer something to the market that is within 10% of intels newest at 1/4th the cost then intel is finished and bankrupt.

>>56152403
>amd's master plan is to create a shitty product and manufacture it for 1/4 the price, a feat that a billion dollar company like Intel cannot do

Do you have a brain

>>56152278

It's beating Broadwell IPC.

The only differentiator is top end Broadwell-e has 10 cores. And a 6900K has a 3.2Ghz Base clock. The engineering sample on Zen had a boost clock of 3.2Ghz. And Broadwell has a TDP of 140W. Zen has a TDP of 95W.

The only thing they can do to fuck this up at this point, is to have such shitty clock speeds that it won't match Broadwell at all. And given what we have seen from the Samsung 14nm LPP in their GPUs it is entirely possible.

>>56152372
But why :c

2500k
gtx770
8gb
256 evo

>>56152278
Are you fucking stupid? Clockspeed wasn't what they were after.

>>56151892
i want it to be true so the goddamn prices can dip lower and force them to actually care.

>>56152733

This.
If this thing is in any way competitive and if AMD starts selling these for ~30% cheaper, then Jewtel has to answer it by practically lowering all of their prices a whole lot.

>>56152495
>Intel cannot do
Do you have a brain or do you simply not understand what a monopoly is?

>>56151068
>Rumor mill has high end at either 6 or 8 cores,

Zen is built around 4 core clusters. If there's a 6 core model, it'll be a fused off binning, like some of the Broadwell-E chips.

There are only 4 and 8 core base variants announced at present though.

>AMD offers an alternative to i7 that has everything you need but Nvidia still has better GPUs

its going to be like the days of my Phenom II/GTX 460 build again
who gonna go AMD+Nvidia?

>>56153581
I might go AMD + AMD just because of this monitor http://www.newegg.com/Product/Product.aspx?Item=9SIAB714894154&ignorebbr=1
I really dislike Nvidia's actions. If they went open, I wouldn't mind them at all.

File: (C88) [Pochi-Goya. (Pochi.)] Ane Narumono An Elder Sister [English] [Facedesk]_001.jpg (148KB, 720x720px) Image search: [Google]

148KB, 720x720px

Zen is great and all, but that 32 core monster though. Holy fucking shit.

>>56153581
While we are waiting for Zen we may as well see how Vega turns out before deciding.

>>56153676
2011
>MOAR CORES
2016
>EVEN MOAR CORES

>>56153658
Funnily enough, Nvidia might actually be forced to implement VESA adaptive sync features in the near future, considering the console upgrades are probably going to be receiving freesync and Intel might implement it for their iGPUs.

>>56153694
Memes aside, it would be fantastic for encoding.

>>56150018
This. Should had used an industry standard PC benchmark tool like SYSmark® 2014.

>>56149949
> LEGENDARY ENGINEER
JIM FUCKING KELLER IS FAR BEYOND THE PUNY LEVEL OF LEGEND

>>56153718
It will win in the en. The problem is that Nvidia isn't caving yet and people still eat up the gsync shit.

File: 070.png (137KB, 481x363px) Image search: [Google]

137KB, 481x363px

>>56150018
>AMDisBETTER_

>>56153718
Open standards always win in the end. That's why openGL exists everywhere and 99% of things.. while DirectX only works on xshit and shitdows

>>56153685
as time goes on and no new leaks happening the more i doubt we'll see vega in 2016 though.
the latest teaser was an AMD employee showing the factory where vega was being manufactured, what did it mean? no one knows.
i wanted to at least see some performance comparisons before deciding on buying a freesync or gsync monitor this black friday.

>>56153745
>people eat up gsync
seriously? theres only like 10 Gsync monitors and theyre all overpriced as hell. I dont think anyone aside from hardcore "professional" gaymurs who gotta have 32 gb ram, intel extreme, and gtx 1999 really care

All the manufacturers are making Freesync instead, and I switched to AMD because of them

>>56149579
>believing tech demos

Bulldozer hype

File: 1471522004927.jpg (10KB, 320x180px) Image search: [Google]

10KB, 320x180px

>>56153772
And the cruel irony of AMD's open source movement is how absolutely abysmal their cards are at OpenGL while they're pretty much *the* brand for DX12 gaming

>>56153809
Shit won't probably last long. They have 2017 for that until Volta comes out. They can always hope they finally ditch Maxwell and build up another Fermi.

>>56153778
Well its not like we have much chance of getting our hands on Zen before 2017 anyways. They are fully saying they don't expect volume production until next year.

So chances are we will get Zen and Vega pretty close to each other.

>>56153804
I want to believe the shitwrecker would only put his name on a product that wrecks shit, not a product that is shit.

>>56153809
VULLKAN

>>56153880
Nothing uses it
Shit tons of stuff uses OpenGL

File: 9oejsp9.jpg (95KB, 550x681px) Image search: [Google]

95KB, 550x681px

>>56149949
So, let's start by playing devil's advocate.

What sort of things did/could have AMD done to this Blender rendering demo to give Intel an edge?

The only thing I can think of is putting a better GPU in the AMD system or use a $1000 Naples Zen chip.

>>56153809
>Who cares if Windows 10 is standard now. If AMD doesn't make their cards run flawlessly on Windows 95, it means they're bad and devote no resources to engineering.

WHAT IS VULKAN, YOU FAGGOT

File: That's bullshit.jpg (260KB, 570x558px) Image search: [Google]

260KB, 570x558px

>>56149949
You forgot your pic.

>AMD stock at 7.04 per share
>tfw you could have quadrupled your investment, but not only were you poor, you never took the risk
How the fuck could I have predicted China? I knew it would climb, but not like this. Not like this.

>>56153936

More likely than not the Zen CPU was using AVX2 instructions to render the scene and the Intel CPU was forced to use legacy x87 instructions. It slows down one CPU massively but still fits under the '''specs''' AMD showed in their demo and there technically isn't a lie :^).

>>56153894
Old software, by nature, is less demanding because it is old. It's not being updated, and it's slowly being outpaced (brute forced) by more powerful hardware.

AMD is putting money in Vulkan, which is the future of demanding graphical applications across the board. Just look at how much better Doom gets when you switch to Vulkan. No, not 10%. Not 20%. We're talking 3x the performance for the EXACT. SAME. FUCKING. HARDWARE.

Vulkan is the future, and anyone who says otherwise either works for Microsoft, or needs to try and steer the dwindling crowd back to OpenGL, which was a failure the second the 2.0 standard was finalized.

File: 1471191376589.jpg (76KB, 640x591px) Image search: [Google]

76KB, 640x591px

>>56153944
>Windows 10
>Standard
I guess it is, whether you like it or not.

but seriously Pajeet, im not going to install win10. Stop asking.

>>56153991
Oh, so AMD is doing to their demos exactly what Intel does with their compiler to real-world consumers.

>>56154026
Windows 10 was an example, you cuck. I know it sucks, nobody uses it. I should have said Windows 7.

>>56151733
holy hell that graph.
what kind of work is FO4 doing on a frame-by-frame basic where main memory bandwidth has any sort of measurable effect?

the amount of engine state needed to control the render loop should be able to fit in L3.

>>56154030

intel stopped doing that 10 years ago, and nobody is holding a gun to your head and forcing you to use icc. anyone competent would just use msvc or clang.

File: tmp_31775-1471577912603698797726.jpg (241KB, 1000x631px) Image search: [Google]

241KB, 1000x631px

BEHEAD THOSE WHO INSULT AMD

>>56154155
Oh shi-, I'm so fucked.

>>56150203
can intel price drop though without severely eating into their spending or without firing a fuck load of people?

Amd can survive on the lower margin parts, can intel?

There is a good chance that this would force intel to put out a gpuless cpu, as that shit eats a shit ton of the die space and then 'drop' the price, possibly shift the cost over to motherboards.

File: intel.jpg (216KB, 1920x1280px) Image search: [Google]

216KB, 1920x1280px

>>56154155
Oy vey goys your greatest ally Intel is here to stop this

>>56153991
um, not a chance in hell, x87 is what runs physx, or at least use to, not sure anymore... there is not a chance in fucking hell that amd is that much slower than intel they needed to do that, fuck, even first gen bulldozer probably doesn't need that much help.

that said, clocking the amd and intel cpu equally, its possible amds cpu doesn't take higher clocks well, so sure at 3ghz its a monster, but intel's out of the box clock makes up for it, potentially the 4.3ghz oc (if i remember right that's a commonly reached 8 core oc for intel) will dominate it.

but realistically, it all comes down to price.

>>56154107
intel did it years ago, but shit still comes out with those tainted compilers.

intel is finished and bankrupt

>>56153426
4, 8, 16, 32

4 is an apu
8 is only a cpu
16 is opteron which is thought o be brought out due to
32 core cpu from a cern leak that said 2 chips on one interposer

at least as far as im able to see.

>>56154273

That's what the Zen+ is all about the following year. More mature fabs, along with some revisions and updating I would love another 10-15% IPC with clock speeds pushing 5Ghz while keeping the 95W TDP.

File: 1470980413975.gif (349KB, 350x233px) Image search: [Google]

349KB, 350x233px

>>56154257
back to your containment board

>>>/pol/

>>56151754
same gpu for each one, there are a few recent games that are heavily dependant on ram. granted ddr4 300 would be good enough seeing as the physics shit themselves over 60fps.

>>56154240
>can intel price drop though without severely eating into their spending or without firing a fuck load of people?

easily.

>Amd can survive on the lower margin parts, can intel?

AMD isn't profitable and they're in danger of being bankrupt in the next few years. They can't 'survive' by selling their products at anything but the highest price possible.

>There is a good chance that this would force intel to put out a gpuless cpu,

intel already sells chips without the integrated gpu

>>56154094
15 year old engine, revamped for oblivion/skyrim, revamped again for fallout 4 and never really fixing underlying problems.

File: GPUs.png (32KB, 789x704px) Image search: [Google]

32KB, 789x704px

>>56154356
A lot of the higher end GPUs have no problem doing 60fps at 1440p ultra

>>56151695
clock for clock, yea, it is marginally better then the 2500, higher clock speed, instruction sets, so on so forth is where intel gets to make its claim of 10% over last year when in reality its so marginally better it doesn't justify the upgrade.

>>56153794
tfw went gsync cause AMD has no 1080 equivalent

>>56154019
of which 4 games use or have announced upcoming use, meanwhile 18 dx12 games out or announced 12 support.

I want vulcan to win and dominate just because fuck win 10 I want to say on 7 but I don't see it happening, at least when one console refuses to let you use it on it.

it would be so great if hiroshimoot implemented ip filtering so i dont have to read the same three or four faggots who shit up every amd thread with imbecilic memespeak

>directx
>>>/v/

>>56154384
yes, and each one of those gpus is likely in a test bench that has the ram maxed as far as it can go too, so you see the game running better on a 980ti then a 960, but pair some shitty ram, lets say my 16gb of ddr2 677 (its 800, but the motherboard didn't like the ram at its rated speed, so it was down clock the ram or overvolt it and i chose downclock because i don't want to fuck with voltage unless its a last resort) and it would preform worse. for some reason, i don't know why, this game is the only one that ram matters that much for.

>>56154483
already could, its on other boards, but for some reason they refuse to roll it out to every board.

>>56154384
thats good to know. too bad that game looks like shit even at 1440p.

File: fallout-4-cpu-benchmark-1440-u.png (31KB, 744x659px) Image search: [Google]

31KB, 744x659px

>>56154550
An FX 8 core/i3 can get 50fps without much issue

>>56152609
Doesn't matter you cuck. They still had to cripple an Intel cpu to beat it.

>>56151874
>buttcoin miners
grandpa pls

>>56154963

You are legitimately too stupid to post in /g/.

And we have a guy posting shit swirls from robotic toilets.

>>56151733
>implying intel didn't steal the design from amd

>>56152265
I can't wait to check out the premium for any of those devices.
>3-5 hour battery life

>tfw waiting for hbm apus

>>56155054

Polaris 11 is utter shit.

>>56153986
buy it now.
it still can go to $20 easily.

>>56152403
AMD priced their way in the lead Athlon 64 line to match that fact. They're never going to turn down what the market will bare.
I'd expect 8/16 Zen to be $500-700 retail, depends on voltage binning
Ht enabled 4/8 prob. around $300
Really sucks there's no 6 core but hopefully it gets figured out with Zen+

>>56151317
>480 will compete with the 480
Wew lad

>>56149949

>node jumps

you sound like a fucking faggot

>>56150018

>unfriendly comments is cheating

W E W BITCH

>>56150682
>>56150127
>>56150998
>>56153730

That's a runtime check you stupid niggers
It doesn't matter what cpu is compiled for, it checks the cpu vendor and cache size at runtime for optimal performance

File: firefox_2016-08-19_08-52-40.jpg (102KB, 1259x688px) Image search: [Google]

102KB, 1259x688px

>>56153721
Yeah you fagotes claim that over and over again...

Yet this happens...

https://www.youtube.com/watch?v=HKXWphkG-a4

Wow pay 500$ more for LITERALLY NOTHING.

>>56153809
AMD's OpenGL drivers may not be quite as good as Nvidia's
But they're still 10 times better than Intel's

>>56151317
But rx 480 absolutely destroys the GTX 480

>>56152239
>implying

fury x
4670
16 gigs of ram

am I poor?

faggot

>>56149579
>on our audiences..... right
He fucking pauses and still gets it wrong.

>>56150018
I'm sorry, but did Zen already lose? Oh, that's right. The 14nm cycle isn't even over yet. In fact, it's only halftime. Does not having the lead at halftime count as a loss? Is that what you're saying? Because if you're saying that I can assure you that you're wrong. Why would you make this post when the cycle is still on? Zen is still being improved right now and it has been the most advanced architecture for how many years now? AMD's competing against one of the most jewish companies in the world who just happens to have a lead because they're feeding off the support of the paying off OEMs to use their CPUs for 20 years. But you know what? They still fucking suck. AMD is one of the most innovative companies in the history of computing, they would be number one if stupid lazy programmers made properly multithreaded software. Maybe you should shut the fuck up before you make retarded posts like this. You know why? Because you're going to be embarrassed when AMD wins and someone bumps this topic. Oh look at that, AMD just released a new CPU that's within 5% of the performance of Skylake for half the price, as many have speculated. Are you a fucking drunk? Are you retarded? Are you autistic? You are a fucking idiot and you should never make a post on this board again and I'm fucking serious. I almost have a feeling you're the only guy making all these anti-AMD topics because you're a faggot hater who doesn't like AMD because they're good. Fuck you, be good at something in YOUR life and then maybe try to troll this fucking company on the board, like I give a fuck. It's so easy to spot out your posts now, you're a retard. Always doing stupid shit like this. Why don't you try to be a good poster? Just for once? For once in your fucking life try not to make a post like this. That's just you, you're always right at getting it wrong. Fuck you. You are nothing.

>>56149949
>believing AMD wont price these on par with the equivalent performing chips
>believing new motherboards will be priced the same as motherboards that haven't had a design change in 4 years

This isn't how economics work. People are buying intel chips at their current prices. They will buy AMD chips that perform the same or slightly better at the same or slightly higher price (if they perform slightly better.)

File: shimomura2.jpg (326KB, 800x591px) Image search: [Google]

326KB, 800x591px

>>56155374
Not a 6800k with 64 DDR4 4 × 1 TB SSD and 5 × 5 TB WD RED with SLI 1080

Yes you are poor.

File: laughing.jpg (395KB, 1500x1100px) Image search: [Google]

395KB, 1500x1100px

>>56155374
>4670
>16GiB

HAHAHAHA

>>56155461
>he fell for the skyline meme
lol

>>56155470
No delete this. 4670 is an entry level processor, as for the AMD "gpu" I wouldnt even use it for digital signage.

>>56155491
>haswell i5
>entry level
yeah, maybe for video editing.

>>56154364
>AMD isn't profitable and they're in danger of being bankrupt in the next few years. They can't 'survive' by selling their products at anything but the highest price possible.
>i know nothing about business
Even VIA and Cyrix are still around senpai, worst case AMD stops being a major player
>intel already sells chips without the integrated gpu
Only on Xeons, this might end up with them rebranding a Xeon into a fuckexpensive unicorn gaming platform

>>56155505
No mate, an i5 is an entry level cpu. If you want real processing power you buy an i7 or a xeon. i5s are crippled i7/xeons anyway. They are like off cuts of the real products.

>>56155513
>workstation cpus
>muh hyperthreading
kys family

File: Screenshot_20160819-102052.png (273KB, 1080x1920px) Image search: [Google]

273KB, 1080x1920px

>>56154410
Kek'd

>>56155513
entry level in what, faggot?

>>56151903
>It's inflated because amd is fucking retarded.
The same has happened everytime AMD launches a competitive product, specially GPU wise due to mining, recall the 290/290X shortages
Though, the RX480 is designed from the ground up to be as cheap as possible, the only reason the price is inflated is due to ridiculous demand, and fucked up distribution channels for yuropoors
>>56154986
Memecoin miners

>>56155523
Your post just reeks of sour grapes. If you want to be poor thats your choice, I will continue posting from my dual cpu E5-2690v4 battlestation.

>>56155513
i cant believe jew tricks work this well

>>56155427
i thought AMD's selling point was their low prices

or am i downs

>>56155114
>I'd expect 8/16 Zen to be $500-700 retail
Then Intel pricecuts i3 to i7's, and launches a ridiculously highly binned Xeon as a EE to keep the status quo
>>56152507
>The only thing they can do to fuck this up at this point, is to have such shitty clock speeds that it won't match Broadwell at all.
It's already running at the same clockspeed as Broadwell
>And given what we have seen from the Samsung 14nm LPP in their GPUs it is entirely possible.
GCN has never been a highly clocked arch, the clocks it is running at are already quite high for such a short pipeline

>>56155558
Yes AMD is only for pajeets, neckbeards, poorfags and furrys.

>>56155326
>Adobe Media Encoder
topfuckingkek anon
Also muh compiling
>>56153794
>I dont think anyone aside from hardcore "professional" gaymurs who gotta have 32 gb ram, intel extreme, and gtx 1999 really care
That's where Nvidia gets most of their earnings outside the HPC market

>>56155513
Nigga what? A fucking pentium is "entry level" if you're looking at pc cpu. I3 would be budget mainstream and the i5 would be high end with all the i7s and above being enthusiast.

>>56155580
why do i have the feeling someone is vigorously typing a hateful reply towards your post?

the reply will have little to no logical sense as well.

>>56150192
AMDs GPUs aren't meant to compete with Nvidia, they are only meant to compete on performance per dollar, which AMD succeeds in.

>>56155580
Do people un-ironically believe this? Intel just sells those scraps so poor people dont have to buy AMD. i5 is entry level. Anything below is destitute poorfag pajeet tier.

>>56155513
>What's wrong goyim? Can't afford an entry level i5? It's cheaper than entry level i7. Who can't afford an entry level Xeon?

>>56155584
Wow you were right >>56155592

>>56151733
I have 800mhz ddr2 and fallout 4 runs fine. RAM speed is the least important spec for a gaming PC.

>>56155607
>(You)ing your own post
You must be desperate

>>56153991
>More likely than not the Zen CPU was using AVX2 instructions to render the scene and the Intel CPU was forced to use legacy x87 instructions. It slows down one CPU massively but still fits under the '''specs''' AMD showed in their demo and there technically isn't a lie :^).
Blender is compiled with GCC, unless they used specially compiled versions for each system both would use the same optimizations, bar the ones that GCC uses specifically for each architecture, which would be less mature than the ones used for Skylake
And you can't know any of that unless you have the binaries AMD used in their test, which no one has
>>56155071
It's probably good enough for a prosumer machine and a shitposting box

>>56155607
>he's trying to justify wasting money on an i7/xeon for gaymen
KEK

>>56155610
>RAM speed is the least important spec for a gaming PC.

Not any more. DF did a video on it and all the games they tested got big boosts when the ram speed was increased.

>>56155636
>Wasting money
$1000 dollars is small change unless you are an AMDtard.

File: ok fag.png (60KB, 1283x663px) Image search: [Google]

60KB, 1283x663px

>>56155618
sorry i rekt you senpai. try not being so predictable.

>>56149579
Stupid question, is this compatible with AM3+ sockets?

>>56155638
Really? Can you link it? Because the fps differences I have found might be 1 to 5fps.

>>56155647
>posting from two devices.

>>56155659
>being this desperate
Lolfaggot.

File: 00ee8b36_FarCry4.jpg (378KB, 1000x563px) Image search: [Google]

378KB, 1000x563px

>>56155652
http://www.overclock.net/t/1586767/digital-foundry-memory-overclocking-and-how-it-affects-fps-in-8-different-games

>>56155690
If you cant afford a i7 6900 you don't deserve to be posting on /g.

>>56155713
I am still running a core 2 quad. So I shouldn't be allowed on /g?

>>56155713
You're in the wrong board, mate. Reddit is more suited for elitist cocksuckers like yourself.

>>56155695
Thank you, these results were surprising for me.

>>56151317
>Sandy lake i3
>Claimed the 480 will compete with the 480
What

>>56155659
>>56155690
>>56155647
>>56155618
Stop it the pair of you. This is /g/. Most of us know what inspect element is.
>>56155659
Go back to /v/.
>>56155690
Just stop.

>>56155713
Why would you want a 8 core CPU in a older architecture? The 6700k is the best for gaming.

>>56155746
Why so desperate for (You)s?

File: dgztfndbdrbDRSB.png (10KB, 370x137px) Image search: [Google]

10KB, 370x137px

>>56155760
What are you on about?

File: k ill stop.png (23KB, 659x381px) Image search: [Google]

23KB, 659x381px

>>56155746
k ill stop

tell me, /g/. if i overclock my DDR3 1600 to try and boost my performance like >>56155695 are they any potential downsides to doing this? eg certain applications not running due to weird clocks or ram dying because of voltage increase.

i have this ram
https://www.scan.co.uk/products/8gb-(2x4gb)-corsair-ddr3-vengeance-jet-black-pc3-12800-(1600)-non-ecc-cas-9-9-9-24-xmp-150v?utm_source=google+shopping&utm_medium=cpc&gclid=CJusk_3_zM4CFUFsGwodhCoMYw

they advertise it as overclocking friendly which i guess helps, right?
>Each module is built using carefully selected DRAM to allow excellent overclocking performance, and has a limited lifetime warranty.

>>56155790
Not as long as you're careful and make sure whatever clocks you end up using are stable.

>>56155638
>>56155652
A bunch of game engines are treating system memory like VRAM, swapping shit to be rendered in and out of it constantly. Huge chunks. Fallout 4 being the prime example, though I believe The Witcher 3 can also substantially benefit from more RAM bandwidth.
AoTS also scales with RAM speed, but thats to be expected.

>>56155748
>gaming
>>>/v/

>>56149949
>literally rendering in blender
>internal demo
>>56150018
and there is this idiot.

t.person who doesnt even own amd product and writing from a Haslel lappy

>>56155961
The guy is complaining that the bench is unfairly favoring AMD's cache hierarchy. Unfortunately for that retard Blender has a feature built in that auto detects and optimizes cache performance. He himself admits that a Blender scene like used by AMD basically only loops through L1.

Hes is something of a well known shitposter, he has a reputation for being incredibly wrong and making an ass of himself.

File: c0f.png (25KB, 640x480px) Image search: [Google]

25KB, 640x480px

>spent $20 more to get crucial ballistix 40nm samsung 1600mhz DDR3 that overclocks to 2133mhz CL9 for my i5 2500k
>thumped how good of an idea it was to spend $15-20 extra for this specific RAM in every build thread and got shut down every time.

SHOWED YOU WRONG /g/, IT WAS WORTH THE $20 YOU FUCKING FAGGOTS

File: 7af6b1091beeaa63656cb45706425bbc.png (11KB, 549x154px) Image search: [Google]

11KB, 549x154px

>>56155790
>>56155804
so i've managed to get the speed up to 2400 mhz stable. shall i push it further or is it not worth it?

>>56149633
Dubs confirm AM4+ wrecking the blue team.

>>56156202
Why don't you try benchmarking video games on your own system instead of asking us?

>>56155695
how can they reach 3200mhz for ram? isn't 2400mhz the maximum on common boards?

>>56156202
RAM overclocking seems to be the most unstable type of overclocking. If you have found something that doesn't bsod on a burn test keep it or keep the number written down.

>>56156291
Most z170 mobos support upto ddr4 3600

>>56155054
With you.

>>56156335
I think he has DDR3 RAM.

>>56156291
Haswell-E and skylake officially "officially" supports ram up to ddr4 2133, 2400 for Broadwell-E.

Unofficially HWE and BWE supports up 3400 and 4200 for Skylake. That is only for 16gb though, the higher you go the lower your chances of getting a good ram oc.

>>56149579
Where were you when AMD defibrillated and successfully revived the Desktop CPU market?

>>56156384
Are we talking 2003? cause I was an 8yo faget back then.

File: datsinglethreadedperformance.png (171KB, 668x624px) Image search: [Google]

171KB, 668x624px

>>56149579
Let me know when AMD releases something that's actually worth upgrading to.

>>56156470
Thanks for CORRECTING THE RECORD.

>>56155910
Most programs max out support at 4 cores.

File: ss+(2015-12-29+at+09.32.51).jpg (54KB, 678x422px) Image search: [Google]

54KB, 678x422px

>>56156360
>Very clearly says DDR4 right there in a huge ass font in the corner of the picture
>I think he has DDR3 RAM.

>>56156470
How does AMD manage to have worse single thread then a core 2 duo?

File: consider suicide.jpg (3KB, 252x44px) Image search: [Google]

3KB, 252x44px

>>56150192
>talking to business journalist about new CPU only
>Nvidia fuccboi shitposts
>Uselessly inflamatory and utterly irellevant
>frogposting

Refer to pic.

>>56156376
>>56156335
T. Hanks.

>>56156593
I own an 8350 and I even admit that bulldozer=faildozer at 3.6GHz, they needed to boost the clocks to 4.0ghz to make it not completely a joke for the money you pay.

>>56156470
>core i5
>being poor

>>56156662
>thinks he has a bigger dig because he owns an i7. Go away AMD peasant.

>>56156715
>owns intel cpu
>amd peasant
logic

>>56156721
>actually thinks hyperthreading benefits you when it comes to gaming.

>>56156738
>implying that it can't
in gta 5 going from my 4.2ghz 4670k to my effectively 4.2ghz 4790k I had a 20-30 fps increase.
plus that stopped the 100% cpu load that would cause the map to stop loading if I drove fast enough.
also remember in the real world you aren't running just the game like all of the reviewers do.
you are going to be running the game + some viop program + a web browser where you will probably have a video playing

>>56156761
I don't go anywhere close to 100% CPU usage. while playing GTA V. Maybe you should remove that malware and viruses off your computer.

https://www.youtube.com/watch?v=ZbQEDaFwFxM

>>56156593
Because that is an 8150 and not an 8350.

Also, according to that, a highly OC'ed 6600K is only 18% better than a stock 5820K in a random MOps./s comparison.

>>56156794
run the game + have an active skype call, and play a 1080p youtube video
I can 100% guarantee you that you will have some pretty high cpu usage

>>56156826
>having that garbage piece of shit Skype running
>playing 1080p videos on YouTube with garbage flash player
How about using some better, lighter programs, such as shit like mumble, discord, and using html5 for YouTube?

>>56156593
>>56156814
IPC in Bulldozer and Piledriver is markedly worse than the Core i arch in C2D and C2Q.
http://www.anandtech.com/bench/product/700?vs=49

With a huge clock advantage you can see the FX 4300 is about equal to the Q9650, each of them showing strong suits in a couple specific areas. The FX 8350 is effectively akin to two Q9650s put together in terms of throughput.
AMD did not make significant gains in perf/clock after K10.

>>56156860
html 5 youtube has higher cpu usage than flash you retard
and yes I only use html 5 youtube.
and ive used those programs before, they are all god awful at blocking out background/unwanted noise.

>Probably cheated
>Probably cheap

God please, I'll take 1 for 2

>>56156860
Already did so. 50-65% cpu usage.

>>56151892
>would be interesting seeing what Intel does if AMD actually is successful.
We already know what Intel will do per history. They will have to compete harder and will likely use unethical means and government collusion to give them more of an edge. That's already happened and it's their playbook.

>>56157006
Its still happening, though now they only tread the line of being legal.
>Sell 10,000 i3s and we'll give you a bunch of free SSDs
>every mobile Skylake you sell comes with a free Atom
>We'd love to give you a bulk discount, but you sold X amount of units last quarter that didn't feature our chips, sorry
>all your laptops have to conform to what we consider an "ultabook" or you get bumped down to low priority for supply timetables

Intel does all of this, even ropes in retailers like BestBuy with their retail edge program. Everything possible to dissuade manufacturers from making AMD systems, and stores from selling them.

>>56157296
Dont forget intel throwing near free atoms with 2gib memory and 32gb storage boards at china with free windows 8+bing to try and get a footing in the tablet and low power markets.

>>56155427
>>56155558
>>56155570
looked this up recently, last time amd was significantly better then intel, they put out a 300~$ cpu that matched intels 1000$ extreme edition, amd then went and put out 3 cpus that were up to 1200$ but were 50% faster then intel.

amd has not been shitty on prices in the past, they likely wont be shitty now looking at the fury x (remember, they had a rage of 550 and 1000$ for the gpu, would nvida, the company that did founder edition undercut the fury x if they could price match it and beat it?) they likely wont be shitty again, this may change for server/enterprise markets, but they intel also puts massively higher margin parts there too. so even if amd was preformance matching, they could still price it 1000$ less then intel and still be several thousand dollars.

>>56155527
comparing those two clock for clock, yea, they would be fucking simmilar, but no, you take an i5 and match it with what intel use to call their extreme editions, no fucking shit at stock, using 6 cores, and 12 threads its going to be better then the cpu that can only use 4 threads.

you have to be being stupid on purpose.

>>56157508
$500 top binned Summit Ridge.

>>56157563
thats what im thinking and would call it fair, however amd could charge sub 200$ for it too, and pull in more money per unit then they do off of their 8350's

hell, for just marketing purposes, this would be genius, get everyone talking about amd, everyone building around amd, get a fuck load of good will from people, and they could do this because they wrote off the pc market as a growth option for them, either in parts sold or profit.

Im in on amd's 8 core so long as its 500 or less, but there is a wide range amd could pull from, imagine if amd put out a cpu that was so good, intel literally could not compete, and then just put out better cpus and charged more once they would outside of intel's performance.

So are all the memory slots DDR4, backwards compatible or only 2?
Does full 8 core Zen have 4 DDR4 channels?

These are questions that need an answer

>>56157663
>amd could charge sub 200$
AMD is there to make money. If they can they'll charge you just like intel. When they had competetive cpus they charged 800+ for top A64FX chips.

>>56152372
post yours anon I fancy choking on my cuppa laughing heartily

>>56157686
Only dual channel ddr4 on AM4, it would be interesting if there was quad channel on am4 for the coming APUs, but then it'd be a huge cost increase negating the benefit.

The upcoming Gxx socket (lga with 2000+ contacts) with 16+ core opterons is expected to be quad channel.

>>56157663
Su said ASP would be higher than their previous CPUs. FX 8350 launched at $195 MSRP.
It will be higher than that pricepoint at least. Margins on their Zambezi and Vishera chips were terrible.

They could price it at intel's mainstream i7, hitting right at $300-$350, but they're not competing against mainstream chips.

>>56157686
Summit Ridge is DDR4 only. Socket AM4 only supports dual channel.

>>56155374
You are a fucking idiot for spending so much on a shit gpu.

>>56157343
I actually took advantage of that with a baytrail tablet that was 80 dollars. Full size USB3.0 32GB emmc mSD card micro HDMI out Win8.1 upgrades to Win10 botnet and plays Crysis. Bought a Zenfone 2 ZE551ML which is a beast of a phone but at the pricing it was sold at I'm sure Intel did some more fudging considering they aren't a thing this year. I love both devices but the facts and reality of the matter seem plain as day.

Intel CPUs nowadays are like cars, just wait a few years and get an older luxury model, saves you a shit ton of money and the end result is the same. The latest generation is overpriced by about 100%.

>>56157726
Each 8 core die has two memory channels, they have an interesting scaling scheme. AMD also promise disruptive memory bandwidth for enterprise.
Memory bandwidth, not cache. Separate points.

One 8 core die = dual channel
dual die = quad channel
quad die 32c/64t = octa channel

8 memory channels with 3200mhz or faster DDR4 is an awful lot for a single socket. A 4 socket blade is going to slaughter in a lot of metrics with memory ops.
If Samsung or anyone else ever releases 3D stacked DRAM DIMMS the enterprise world is going to eat it up. Several TB of stupid fast RAM is about as good as it gets until Hp materializes their memeristors.

>>56157736
Is it really a horrible GPU considering it's cheaper than a GTX 1070 and completely demolishes it?

>>56157726
Thinking about it I can see how two channels of DDR4 fits the design if that 100GB/s core complex interconnect is true.
38GB/s or more from the IMC and headroom for L3 traffic simultaneously makes sense.

Kinda worried about clock speeds though, I fear having optimized for power and density will make max clocks suffer 10-15% compared to Intel counterparts. Feels like Phenom II all over again.

>>56157763
That means more and more sockets though. There must be a limit to the inter-die interconnect (HT4?) as well. Unless there's another use for the imc which can't connect through the socket such as memory on an interposer, unless that's what you meant with the stack dram.

I don't actually know what disruptive memory bandwidth is yet.

>>56154873
>A FX9590 with 4,7GHz on par with a stock haswell i3

HAHAH goddamit, why would ANYONE buy a AMD CPU?

Zen will be a failure, Vega will be a failure, mark my words.

>>56157794
We won't see on-package stacked DRAM on CPUs until yield, capacity and bandwidth requirements both skyrocket.
With many HPC systems using 64-128GB per board they'd need at least 32GB of HBM or equiv. on package and a robust L3 cache to hide typical stacked memory latency

>>56157825
(you)

I'm really interested in seeing how new APU's compare to say the 460 card itself. If they can get near that performance I might actually ditch the Roku's and build around AMD APU's in HTPC's instead despite the cost. My other concern there is that typically AMD isn't great HEVC H.265 encodes but we'll see. I really want to move to H.265 to get those file sizes down while keeping the quality.

>>56157794
wot
A single socket can hold a 4 die MCM.
A 4 socket blade would give you up to 128 cores and 256 threads.
Each of those sockets having 8 channel memory.
With 3600mhz DDR4 thats 230.4GB/s per socket.
Thats what they mean by disruptive memory bandwidth.

Current Xeons, including the big 24 core E7 8890v4, are only quad channel.

I meant exactly what I stated about TSV DRAM:
http://www.extremetech.com/computing/192711-samsungs-new-20nm-ddr4-clears-the-way-for-massive-128gb-dimms

Extremely high capacity DIMMs, coupled with extremely high bandwidth makes for incredibly high throughput in memory ops.

>>56157716
I just mentioned the period where they did beat intel ee 1000$ cpu with a 300$ one right? did i forget that? no, >>56157508 there it is. look into it, was fascinating to me seeing that because everyone screams 'amd would be as much of a jew if they could be' but when they could they weren't.

lets not forget, to investors THE PC MARKETPLACE IS NOT A REVENUE GROWTH AVENUE FOR THEM

you can interpret this as either they are planning to not sell as many parts, but the more likely one is they wont be making more profit then they do now.

amds current cpu the 8350 came out at 200$ and is currently sold at 150$ and the 8 core zen is literally half its die size. they could charge the same 150-200$ and pull in more profit per unit. they could sell it that low and pull in volume, they could sell it that low and get people talking about amd and building around amd, and when zen plus comes around up that cost to 250-350 and above, because they completely wrote the consumer market off for growth revenue.

>>56157735
again, look above, last time intel competed with intel was with the extreme editions, what is what the 8 core and 10 core are just by different names now, and amd charged 1/3 the price, and proceed to sell the ones that were better at the marked up prices.

you have to look at this as more then profits, look at sony and microsoft who are willing to sell things at a loss, amd selling the chip cheaper would be marketing on amds part, they could also do the balancing act of lower price higher sales vs higher price lower sales.

personally I think the thing will be i7 priced, possibly i5, but im prepared for 500$.

>>56157851
>amd
>ddr4
>ddr4 3600
>ecc ddr4 3600

pick one

>>56157851
no, the 32 core 64 thread server behemoth they showed was a dual socket system
Their high end server platform is a 16 core MCM, two 8 core chips plugged into one board

The only 32 core AMD product was a fuck huge APU with HBM 'heavy on the salt' rumor which may well be fake or just wishful marketing.

>>56157851
I was very clear sorry.

What I mean by more sockets was something like a 4die 32core sku would have 8 memory channels. If it's going in the same socket it'll have four of them wasted unless they can put them to other use within the package, some kind of l4 memory for example. Unless they make another socket for the larger chips which is probably uneconomical, as you said the larger xeons are still on 2011 with quad channel.

>>56149579

File: Zen_Summit_Ridge_First.jpg (73KB, 369x676px) Image search: [Google]

73KB, 369x676px

>>56157886
All Zen based parts have a DDR4 memory controller
AMD detailed their new Opteron platform as supporting DDR4 3200

Shitpost somewhere else

>>56157896
16 cores is two dies.
32 cores is four dies.
That is how the GMI links work.
No one is talking about an APU here.

Stop with your stupid guesswork.

Its a good teaser as it has sent a lot of idiots into a frenzy of shitposting across the various tech sites around the web. Apparently some of these armchair engineers have already deduced the entirety of zen's performance based upon this benchmark which doesn't even state hardware composition beyond 1) they are equal and 2) both chips clocked to the same frequency.

>>56155421
rip anon who got burned

Hail AMD

>>56149579
i bet its as legit as the 480 leaked benchmarks

>>56157913
There is no 4-chip MCM and I will not post further beyond that.

>>56158090
Yes, there is. That is how they achieve 8 channel memory. A 4 die MCM, two channels per die.
Stop talking out of your ass.

File: dkzh.png (35KB, 180x198px) Image search: [Google]

35KB, 180x198px

>>56156470
this is absolute bullshit, when i ran this test on my phenom II, it listed the 955 there as being over 1300 and ahead of the core2 duo. The numbers on the baselines changed when i switched to an intel cpu and didn't update passmarks program. They literally have 2 sets of baselines because 2 different workloads depending on your cpu manufacturer? Fuck them.

How can a Phenom II 955 be slower than an e8400 in single threading? It's objectively faster at int and float. Fucking benchmark favoritism.

>>56158228
Passmark is not a valid bench.
Thats why. Never has been and never will be.

>>56158228
>tfw i almost got memed into buying a phenom 2 and a 5850
>mfw i got an i7 instead with a 470, which i rma'ed earlier this year for a 950

File: 739228bc3981c1a419c8c5480d5baf42.jpg (10KB, 480x480px) Image search: [Google]

10KB, 480x480px

>>56157884
>'amd would be as much of a jew if they could be' but when they could they weren't.

Daily reminder
https://web.archive.org/web/20130719043005/http://www.newegg.com/Product/Product.aspx?Item=N82E16819113347


and this didnt have near the amount of hype as Zen, hell it was just an oc'd 8350. Are you starting to see now? Every company is greedy. Just because AMD pushes open source/standards doesn't mean they're out to give us as much as they can for our money.

>>56158270
>I have 1060 equivalent card
>I have RMA'ed it for 950

Yea, Believable.

>>56158299
the 470 was a 1070 equivalent, difference in performance 470 vs 480 back then was virtually identical to the 1080 vs 1070 now.

What is the point of new cpus when the 2500k still crush everything we give to hum ?

It doesn't matter how well amd performs, nvidiots and jewtellians will always, ALWAYS insist that because they're paying more, they're getting inherently better performance, and show benchmarks from a Nvidia goyworks game. And they will ALWAYS use obsolete arguments, such as "what about that one time?" Or "what about this one benchmark?" This is why I never argue with them, no matter how well you make your argument, it's wasted.

File: certified shit wrecker.jpg (13KB, 381x286px) Image search: [Google]

13KB, 381x286px

>Summit Ridge dies are just Opteron candidates that didn't bin at the right clocks and voltage
>all enterprise parts have the exact same layout
>each die has two GMI links and dual channel IMC
>4 dies when oriented correctly all have their GMI links facing one another
>Papermaster explicitly spoke on providing disruptive DDR4 memory bandwidth
>linux kernel patches show socket SP3 supports up to 32 physical cores
>leaks have explicitly shown 32c/64t Opteron parts
>intel right now is offering 22c/44t Xeons
>this retard ITT thinks AMD won't sell 32c Opterons

It doesn't get much dumber than that.

>>56151903
You really think amd cares about what a few NEETs on g think?

>>56158299
>>I have 1060 equivalent card
lol in name only. performance wise a 950 shits on a 470.

>>56158267
>>56158290

that is an interesting one, I honestly have no idea what the hell that even was, wiki tells me that it was msrp at 270$, 3 months after release it was 300$, and this part was only ever an oem, so I have no idea what to say...

you look at one thing, and say they are going to be shitty, Im looking at everything they did and see a general trend to not be shitty.

>>56154474
You can't compare the two APIs based on the number of games using the two APIs when DX12 became available to the public with Windows 10 over a year ago when Vulkan has been out for barely 6 months.

These days few developers build their own engines, instead relying on third party solutions and we know that at least Unreal Engine, Source and Unity are ether in the process of implementing Vulkan support or have already done so. Then there's also developer internal engines and we know that id Tech (Bethesda's internal engine) already supports Vulkan while Frostbite, EA's internal engine, is in the process of moving to a Vulkan render.

>>56158417
you might be thinking about 460
470 is about 10% worse than 480.

if you are thinking about 460 then getting 950 is correct action as it OCed to 960 levels

if you had 470 then you have fucked up royally because it OCes to 2% less than 480 OCes to.

>>56149579
>delayed to 2017

Plenty of time for Intel to catch up, really.

There has been a lot of suspicion that Intel is artificially keeping CPU performance low because there is no competition thus no reason to release higher perf processors before the 24 month upgrade cycle on high-end PCs out there are up.

>>56158442
>if you are thinking about 460 then getting 950 is correct action as it OCed to 960 levels
How?
http://gpu.userbenchmark.com/Compare/Nvidia-GTX-460-vs-Nvidia-GTX-960/2167vs3165
It's ridiculously slower.

>>56158437
>These days few developers build their own engines, instead relying on third party solutions and we know that at least Unreal Engine, Source and Unity are ether in the process of implementing Vulkan support or have already done so. Then there's also developer internal engines and we know that id Tech (Bethesda's internal engine) already supports Vulkan while Frostbite, EA's internal engine, is in the process of moving to a Vulkan render.
we also know unreal 4 was built ground up for gameworks so good with the bad?

but seriously, they had 6 months to announce support, likely longer then that to start working on vulcan as they likely could have had access to non finalized api. it honestly looks like vulcan is dead while dx12 has a fuck ton of announced support and the obvious microsoft support in porting their first/second party to dx12 only.

>>56158444
>Plenty of time for Intel to catch up, really.
Not with time-to-market times of modern CPUs... We're talking something like 4 years from finishing principal design and over a year from tapeout (i.e when the design is finished and sent for production).

I really doubt that Intel has a finished microarchitecture ready for mass production just lying around in case AMD suddenly catches up because that's the only way Intel could somehow suddenly take a huge leap forward.

>>56158509
Intel may even have some ready to ship in their warehouses.

That's how much they've delayed releasing new products to clear out the inventory.

File: Screenshot from 2016-08-19 16-15-53.png (115KB, 1142x541px) Image search: [Google]

115KB, 1142x541px

>>56158472
I hope this is just parody...

>>56158472
You son of a bitch, Fucking hooked me up and reeled me right in.

Well fucking played.

>>56158442
a 470 overclocked to 850mhz matches the stock 480 or a 560 ti 448. the most modern card with the performance of a 480 is a 750 ti.

a 950 overclocked will match a stock 680 or a 770.

it takes two 470 overclocked in sli to reach stock 680 performance, that's not accounting for the nearly 650w power draw of the two overclocked first gen 40nm cards. so to get a 950 for a 470 is a pretty fucking good deal. the lifetime warranty carries over too.

had i gotten the 5850 the warranty would've expired long ago because no amd aib offers lifetime warranty and not only that, amd pretty much abandoned the 5000/6000 series in 2013, or 2015 offically.

>>56155421
this nigga went in

>>56153740
this

>>56158475
>but seriously, they had 6 months to announce support
They were on the list of companies involved in the API's development well before the final spec was released and the official announcement that UE would support Vulkan happened less than a week after the spec was published in February this year.

Like DX12, developers obviously won't start announcing official support for games right away so we're just going to have to wait to see how things go. However try to remember that Dice has been hesitant about confirming that the PC version of Battlefield 1 will use their DX12 render.

>>56158518
>Intel may even have some ready to ship in their warehouses.
>Being this delusional
Just no... Intel doesn't have any extra silicon lying around as some sort of "Break glass if threatened by AMD" measure.

>>56149949
I hope Zen is like the AMD Athlon XP of 2017

>>56158603
>Intel doesn't have any extra silicon lying around as some sort of "Break glass if threatened by AMD" measure.

They have a warehouse full of Kaby Lake already.

What's it for?

>>56158625
>Implying
No... Only reason why they'd have Kaby Lake chips lying around en masse would be because they'd overestimated demand and made too many of them, not as some kind of safety guarantee against AMD.

File: 1440674318077.jpg (84KB, 750x709px) Image search: [Google]

84KB, 750x709px

>>56153986
Fucking this , if i actually was not a poorfag NEET i would have bought like 200 of them while they were 2$ a share.
>was planning on maybe getting them later because zen was late 2016
>mfw china happens

I just want a 35-40w CPU that's at least as fast as my 3570k, and an itcx board to put it into.
Here's hoping.

I'd also like it to actually be available to consumers at launch, unlike intel's low tdp parts which it seems you have to wait until people start pulling them out of old servers.

>>56158342
>search 32 core zen mcm
>top four articles are wccftech
>next 4 are linkbacks through every other shitty rumour mill
>this one anon is retarded

>>56159384
Yes, you and that anon are both retarded.

>>56153804
This. Shaping up to be Bulldozer 2.0.

>>56159384
http://www.anandtech.com/show/10581/early-amd-zen-server-cpu-and-motherboard-details-codename-naples-32cores-dual-socket-platforms-q2-2017

At the AMD Zen microarchitecture announcement event yesterday, the lid was lifted on some of the details of AMD’s server platform. The 32-core CPU, codename Naples, will feature simultaneous multithreading similar to the desktop platform we wrote about earlier, allowing for 64 threads per processor. Thus, in a dual socket system, up to 128 threads will be available.

32c/64t CPU
A dual socket system having 128 threads.

>b-b-b-b-b-b-but there won't be 32c CPUS!!!

Tech. Illiterate. Retard.
You're probably some 80 IQ third worlder too.

Actually they have been cash flow positive since last earnings were posted

File: 205.jpg (30KB, 542x616px) Image search: [Google]

30KB, 542x616px

>>56150192
I want to fuck Lisa Su

File: 1470768184397.jpg (64KB, 1037x784px) Image search: [Google]

64KB, 1037x784px

>>56149579
>it's legit? It's not a scam like the rx480 cf?
They did say CF 480 beats a 1080 in DX12
>It actually does
Now i believe this.
GO GO AMD
Zen Vega is gonna be good!

File: shit wrecker.jpg (62KB, 800x400px) Image search: [Google]

62KB, 800x400px

>>56159384
>>56158090
How can you be this clueless?

>>56157825
Obviously that 9xxx series chip wasn't stable

>>56159612

Gamersnexus tested it on the msi 970 gaming motherboard - a board not rated to handle a 9590. You'll find a lot of sites test the 9590 on boards not rated to handle it.

>>56159650
Steve's personal rig runs a 9xxx series chip, and I think not even his 990FX higher end board runs it stable

>>56159671

Only about 5 990fx officially support the 9590 (most of them from asus).

File: s1 Perf.png (266KB, 2000x1100px) Image search: [Google]

266KB, 2000x1100px

4 core APUs scaling from 4w to 95w
8 core desktop chips
32 core servers

Incredible scaling for a wide core. AMD is detailing the arch in full at HotChips in just 4 days.

>>56160361
The more I head about it the more I know I will by this on day one.
Seems like real winner.

>>56160388
>buy this on day one
Don't be a retarded fucking consumer you fucking retard. Wait for the reviews. They look good, by all means buy it.

File: wyoming36.jpg (1MB, 3264x2448px) Image search: [Google]

1MB, 3264x2448px

>>56160388
I was thinking of holding out for Raven Ridge and just using that to replace my dinky web machine/HTPC.
I've been more and more tempted to build a decent video editing system though. I travel a bit, and I've been considering starting a youtube motovlog channel to try and get some of that easy adsense money. A 65w 8 core Summit Ridge would be pretty nice for that.

>>56160388
Kek. Did you buy Bulldozer on day one too? You should listen to this guy >>56160409

>>56160579
No I used only intel so far but i need something on a budget that can perform and this looks good.

>>56151317
b8

>>56155421
>Oh look at that, AMD just released a new CPU that's within 5% of the performance of Skylake

True

> for half the price.

That we don't know. I will upgrade my 5820k to 8/16 Zen at proper clock in a jiffy if it proves a sideways upgrade tho. I will gladly make someone pay intel tax and get two more cores.

>>56155512
HEDT cpus don't have IGPU as well. Altho you might argue they are xeons as well.

>>56157884
>THE PC MARKETPLACE IS NOT A REVENUE GROWTH AVENUE FOR THEM

Pc nowadays is strongly tied to the 'next big thing' that will require lots of CPU horsepower. If we get games that spread over 8-12 cores out of the box, then PC will start to be a growth market again. So far we are stagnant in catch 22 situation, where devs don't invest what players don't have, and those don't invest into what they don't need.

With vulkan/dx12 spreading nicely over multiple threads we just might see push for newer, multicore ( 8+ ) cpus.

File: trumpet.jpg (26KB, 500x633px) Image search: [Google]

26KB, 500x633px

>>56150018
Piednoel is quite the clown. Funny how wrong he managed to get with those tweets, he even had to delete them, tchhh.

File: 1430277802068.gif (2MB, 480x270px) Image search: [Google]

2MB, 480x270px

>>56152178
>>56152192
>>56152239
Honestly, these /g/ "image" creations are shame of 4chan. Give it some effort at least, you kiddos. Come back when you can do something worthy like this.