AMD 7nm Vega 20 32GB GPU 3DMark Performance Benchmark Leaked

wccftech.com/amd-7nm-vega-20-32gb-gpu-3dmark-benchmark-leaked-up-to/

videocardz.com/76076/early-amd-vega-20-3dmark11-benchmark-result-emerges

Attached: 1510092879994.png (900x600, 492K)

Other urls found in this thread:

anandtech.com/show/12677/tsmc-kicks-off-volume-production-of-7nm-chips
twitter.com/NSFWRedditImage

>Vega 20 70% Faster Per Clock vs RX Vega 64 Liquid

#team red

NVIDIOTS BTFO, GREENFAGS ON SUICIDE WATCH

It's a GPU without a display output, not for fucking games

>not for fucking games
No shit it's amd, so it's probably only good for in some obscure use case that isn't games, or most professional use.

I just want a low end Vega or Navi to use with Linux while I passthrough my higher-end GPU to a VM.

surely cause its fucking impossible to create another gpu with the exact chip but with an video output

Jow Forums everyone

Frontier edition with MI25 bios can also drive a display, so support is there in principle.

If this is true then it seems like AMD was able to push clocks significantly and that 1000Mhz was only a wrong reading. That and some fixed in-silicon bugs.

>j-just wait guise
>THIS time nvidia is finished
amdrones never learn, do they

its an eng sample user we arent going to see those card for quite some time obviously they are testing things..

>2nd half 2018 release
It's not 7nm.
No 7nm node is ready for volume production until q1-2 2019

Nvidiot amnesia at it again
Don't you remember Fermi? GeForce FX?

its just AI card who cares
this doesn't look like a gaming video card given it has 32GB vram

This.
Every single time we have these thread spammers talking about muh AMD going to win this time guise.just wait for it.
This will be vega frontier aka block chain pioneers all over again.

Attached: ayymd.webm (1280x720, 1.23M)

those cards arent for volume production user they are made with hovis method for a very short volume

That was a crazy time for AMD, literally 4 generations in a row winning at power consumption.

SHUT IT DOWN

>TSMC last week announced that it had started high volume production (HVM) of chips using their first-gen 7 nm (CLN7FF) process technology.
anandtech.com/show/12677/tsmc-kicks-off-volume-production-of-7nm-chips

Stay mad butthurt intel kike shill, your shitty process is a disaster.

>per clock
What to expect? 1.28 TB/s sound tasty

I have a 1080 Ti, but I would not mind buying a ice Sapphire custom card for a reasonable price
But I won't buy anything that doesnt deliver at least 50 % more

The FX line hasn't been used in over a decade.
The gtx4xx series wasn't bad at all. Sure it was beaten in some benchmarks, but when it came to newer features like tessellation it destroyed Radeon 5xxx series. In fact the gtx 480 even beat out the 6970 in some of these tests. Even so, AMD hasn't even competed since the gtx 7xx series.

Also don't tell me about vega. The vega 64 is outperformed by a regular 1080 that was released over a year earlier and uses half the wattage.

Not an "nvidiot" btw. I have a 6970 and 2 r9 290s.

>Slower than POOGA

TOP KEK, AYYMDPOORFAGS

I remember how horrible the 290 series were on release, compare how they are now. A 280 sometimes beats a fucking Titan, the 290 was on par with the 780 back then. Now its faster thn a 1780 Ti.

Its still a good card, xcept the power draw, it competes current gen midrange cards.

GTX 1060 120W outperforms 290X/390X 300W HOUSEFIRES while providing features that people want, DisplayPort 1.4, HDMI 2.0b, HEVC Main10 & VP9 10bit 8K hardware decoding

AYYMD HOUSEFIRES, LITERALLY THE IDIOT'S CHOICE

Whats with this fanaticism over some brand?

I swear to god nvidia is paying people to post this shit because who else would want monopoly.

they're called nvidiots for a reason

AYYMDPOORFAGS fanATIc, everyone

>OP nvidia wojak
>WAHHHH WHY IS EVERYONE SHITTING ON MUH AMD
kys

>half the die manufacturing process
>38% slower gpu
>26% faster ram
>twice the ram size
>10% slower overall
why bother? is this for laptops or something?

So is this when AMDrones switch from "wait" to "hype" mode again?

here we go again

That GPU is a test part fabbed at TSMC, and TSMC is in fact ready for volume on their 7nm line.

The moar you know.

you know, after purchasing 1070 I feel calm about all this going in rounds
but you know even the blind hit bullseye sometimes, one of these days AMD will deliver

There's 3 versions of GP100.

P100 Tesla PCI-E sans display port ( and lower clocks )

Quadro GP100 PCI-E ( same as above but more expensive )

Quadro GP100 PCI-E ( 3/4th memory )

>GTX 1060 120W outperforms 290X/390X 300W HOUSEFIRES

GTX 1060
released July 2016

R9 290X
Released October 2013

R9 390X
Released June 2015


Wew, who knew something several years newer would be better. I'm shocked.SHOCKED i say.

How new are you? Vega was heavily shilled here for a fucking year, then turned out to be the most disappointing launch in recent GPU history. Hey AMD, where's the primitive shaders you promised? Oh wait, it's up to developers. (Meaning no game will ever use it.) Literally ANY criticism of Vega was meant with people spamming the whitepaper, which ended up being full of lies and missing features. Fuck RTG and King Pajeet, Vega was literally a scam GPU.

Remember when Vega was meant to destroy Pascal's successor?
Good times.

Attached: bVobYT.jpg (1188x909, 489K)

They just added more stream processors, that's something nvidia can do as well. Don't believe for a second that any gcn based card can compete with anything nvidia offers (especially considering performance per watt, remember that when amd makes the jump to 7nm so does nvidia and what can cut it half vegs 290W can also cut in half pascals already lower consumption).

King Pajeet left for Intel.

Maybe 2020-2022 but not as long as gcn is around

>he thinks it's based on GCN

Attached: FB_IMG_1524974626951.jpg (340x962, 47K)

It is GCN, but it's the first 6th generation GCN part, so comparing it to GCN 5th gen is probably retarded and has little to no bearing on reality.

according to AMD when they introduced it last year it runs on the successor to GCN? maybe i'm misremembering

The marketing guy who was responsible for this is now works at intel together with Raja.

Wiki has it listed as 6th gen GCN.

Could be something else though, wont really know for sure until we see actual samples going out to reviewers.

I quite honestly don't care what new GPUs AMD and Nvidia are releasing right now, not even a little. The only actually important GPU related news is the fact that Bitmain is finally shipping those ETH ASICs. New GPUs mean jack shit when even mid-range GPUs are priced like it's a whole computer. Nvidia's 1060's priced like it's a 1080, AMDs RX 580 - 4 not 8 GB version - is priced like it's a Vega64.

Attached: hot-beautiful-kwon-yuri-3.jpg (467x700, 44K)

Just 24 hours after leaving AMD, Intel hires Raja Koduri as lead GPU architect
Intel hires AMD product marketing veteran Chris Hook

Russian man sentenced for killing friend in AMD vs. Nvidia GPU dispute

AMD Annoints Two New Heads Of Its Radeon Technologies Group
>Now AMD has selected two industry veterans to lead RTG.
Mike Rayfield is now the senior vice president and general manager, and David Wang is now the senior vice president of engineering

better than Raja?

Attached: 1d.gif (690x200, 973K)

Vega is still every bit GCN.

NOOOOOOOOOO!!!!

Attached: 1521563380669.png (1300x599, 305K)

The second one can display output unlike the Tesla version ( saves on cost )

I like nvidea, up until they added spyware

you dont have to install geforce experience
there is also slowing down older cards and
Geforce partners program
>muh gaming name
stupid but it sells.

Here's an interesting question: What happened to VEGA12? AMD seems to be doing some kind of low/mid-range line since support for a VEGA12 chip (keep in mind that Vega56/64 is VEGA10) was added to the Linux kernel a month ago. This new rumored card is VEGA20. I've seen no VEGA12 rumors on sites like wccftech and no mention anywhere. By the PCIID's it looks like a whole range of cards. I do wonder.. perhaps both nvidia and amd card sales dropping something like 40% the last month (due to crypto ASICs) made AMD delay or drop that line?

The simple truth is that Vega was a massive success for AMD. Yes, really. I know it's because miners, not gamers, ended up buying every last one even though the prices have been and still are wildly inflated. If you ignore gaming performance and this and that and just look at the raw sales numbers.. it's a huge success. Yes, I know calling it a success is .. odd.

What happened the last time? It was exactly the same with Vega. Was supposed to change the game, ended up being more expensive than a 1080ti and performing worse than a 1070ti.

Don't let the pajeets tell you it's good before it's released.

Attached: 1523057059520.png (1080x1920, 838K)

repeat after me
1070 level performance

Spyware also gets installed with the Nvidia driver only
It can be uninstalled/deleted afterwards though

>nvidiot here
I'm only glad for some competition.

Volta still MIA.
Perhaps AMD marketing wasn't wrong?

its just user id and such or no?

>Nvidia doesn't bother releasing the Pascal successor because nothing from the Poolaris or Vega lineups can touch the high end cards years after release
>YES, WE WON!!!!!1 XDD

Attached: 1524956317845.png (1296x1458, 217K)

>tfw you wanted a Vega but nvidia cards are starting to become a bit cheaper

NVIDIA's a scummy company and doesn't provide hardware documentation like AMD does. This is even after they drop support, which they tend to do quite early. Their cards also tend to break after just a few years of use. And NVIDIA's support will just tell you to "upgrade to a new card" if you contact them for any reason and your card isn't current-gen.
Due to all of this, I got a Nitro+ Vega64, and I'm glad I did.

how many MHash/s?
buying day 1 before miners.

brian justnich on suicide watch

telemetry

3DMark11? wtf???

vega 64 is pretty close to 1080 and vega56 wins against 1070. AMD only failed at wattage, timing and in that they dont have a 1080ti competitor.

Overclock it.

Attached: 290x overclock 08.04.18.png (1275x691, 497K)

Its a 3dmark score that leaked.

Fuck off.

That's great. But then you have to wait years for working drivers.

Attached: 1795t4.jpg (500x616, 81K)

NOOOOOooooooooo

I have both cards and the 290x is faster

Ever pushed your 290x to pull as much juice as vega?

I've only had this overclock pass unigine valley once. These days I keep it at 1160mhz core because I don't need to feed the core 1.4v which is nice.

Attached: 290x biblical overclock on air.gif (537x781, 21K)

Bs. The 390x is way faster. I have both.

Noice will look great in my crypto farm

t. intel

>we are already back to the "t-the n-next a-amd thing w-will k-kill n-nvidia I p-promise g-guyze!" overshilling phase

wew, it literally never changes indeed

Attached: 1438392446246.png (361x500, 187K)

release fucking when tho?

Flip it the other way - Jow Forums has declared AMD dead and bankrupt FOR REALS GUISE every year since 2011 (and probably earlier).

absolutely atrocious afterburner skin
use classic you moron

>Buying buttmining cards

>mine on idle time
>game on gayming time

Why should I care what it looks like? It spends 95% of its life minimised in my taskbar.

yeah but it looks horrible when it's not

I have been out of gaming for a while is a AMD R9 270x at 1.1Ghz still able to play some games like doom 2016?

Or do I need a new card?

>32GB GPU
Why is this necessary?
I've never even used all 24GB of RAM on my machine.

LIES ALL FUCKING LIES FUCK YOU I DON'T BELIEVE YOU NONE OF THIS PAJEET SITE BULLSHIT IS TRUE IT'S JUST ANOTHER WAITING GAME FROM AMD TO STOP PEOPLE FROM BUYING SUPERIOR NVIDIA GPUS FUCK YOU FUCK YOU JUST FUCK YOU

What do I upgrade to from a 380? It's holding pretty well desu but just want to max out 1080p with AA more at 60fps. Is the 580 worth it? I've noticed the x80 series got a lot more expensive, does it still cover the same market segment as the 380 at launch or higher?

>doesn't provide hardware documentation like AMD does

Buuuuuuuuuuulshit, amd are terrible

>features that people want
>gsync
No one wants that shit. They're forced into it.

I'm seeing 10% faster.

It's GCN2, basically. Polaris was 1.4 or something. It's not the exact same, but it's not entirely new.

It's not really worth it as an upgrade from a 380. A 380 is still really good for 1080 gaming. Better than a 1050Ti by far.
Only reason to upgrade would be lower power usage. Just wait for Navi senpai if Vega 7nm isn't coming to consumer.

>there are illiterate plebs who don't even know the radeon instict 7nm vega was always meant to be a deep learning/AI chip

Attached: 1302473028612.jpg (580x303, 30K)

Holy shit. At this rate, how much longer until we get to see a demo of what Zen 2 is capable of?
>TFW still using a processor made on a 65nm process
>TFW next years there will be as many die shrinks between processors currently being made and my current processor as there were between my current processor and the P54C Penitum/486DX4

>32GB HBM2

It's a firepro card

Sometimes I wish I had got 285 instead of 960 2gb
It's also incredible they kept it competitive enough for such an old 7000 series card

>I'm seeing 10% faster.
you might need reading glasses

Attached: AMD-Vega-20-vs-Vega-Frontier.jpg (1081x1340, 121K)

It's also clocked a shit ton lower

i dont really like to repeat myself, so see

Attached: image.png (662x445, 40K)

From the wccftech article:
>Now, the most interesting part here is the GPU clock speed. If the reported figure is to be believed this early engineering sample of Vega 20 is running at exactly 1GHz. It’s possible that this is because 3DMark is misreporting the clock speed. However, if we go back to really early Vega Frontier Edition database entries we find that when 3DMark fails to read the clock speed it simply reports it as 0, rather than 1GHz. And misreadings of the UEFI data have also led to misreadings of the HBM2 vRAM by 3DMark. We don’t see any of that with this particular Vega 20 entry

>there are illiterate plebs who don't even know the radeon instict 7nm vega was always meant to be a deep learning/AI chip
>AMD
You mean that the 7nm vega was always meant to be a not as good "obscure niche use case" waste bin tier chip.
People who buy this will be paying AMD money just to throw it away for them.