Same device ID as the flagship Vega sample demoed a few months back too

Same device ID as the flagship Vega sample demoed a few months back too.

Good job AMD, first you show your incompetence with Poozen then royally fuck up Pooga.

Other urls found in this thread:

3dmark.com/spy/1544741
videocardz.com/amd/radeon-instinct/radeon-instinct-mi25
anandtech.com/show/10905/amd-announces-radeon-instinct-deep-learning-2017/2
twitter.com/NSFWRedditImage

t. angry corelet with a shitty i7

>rumours
kek fucking maximum damage control kikes

>BUTTMAD AYYMDPOORFAG DETECTED

Look at the device ID in the Doom image, and look at the device ID from scrolling down

JUST WAIT
U
S
T

ayy

>that image
saved

I'm already saving up screenshots for new OC coming up with the release of Ryze. You subhumans never learn.

DELETE THIS NOW DELET THISSS

As always

I can agree, they fucked up with Polaris and with hyping up Ryzen as the best gaymen machine. But this is just bad. Ryzen 5 is now the only good thing AMD has done

Little Vega, remember.

>Little Vega

Big vega is dual gpu isn't it?

I can already see it. "1.5X greater performance than GTX 1080"

>gpu clock 1200 mhz
into the trash it goes

Rebranded Fury X at14nm with 8GB HBMeme2 ?

>1200Mhz
>700MHz

it always lies on the price point. it could be literally the next shoa for nvidia

the last leak was @ only 1000 Mhz

That's not a leak, that's a placeholder page, with completely wrong details on top of it.
Wrong GCN version, wrong bandwidth, wrong clocks, heck it even said April 28 release date a few days ago.

>performance estimate based on arch, shaders and clock

3dmark.com/spy/1544741

>10.5 inches/267 mm

That very accrute, I don't think it's bulshit.

>inb4 hurr durr techpowerupshill

That's the size of a regular Nvidia/AMD aftermarket card, we've already seen images of Vega and it's the size of Fury X.

Also arguing as if the fucking (wrong) dimensions which have been standard for years already somehow makes the technical specs written there not valid is beyond silly.

When did this board turn into Sup Forums-tier console war garbage?

So it's 5% faster than Fury X? That makes zero sense, we're not talking a 28nm GPU there.

Vega has ~40-50% more transistors and 45% higher compute performance than Fury X.
If you believe a benchmark that shows Vega is only 9% faster than Fury X, then you have shit for brains.

You retards do realize that this can't be real, right? They could've overclocked a Fury X and gotten the same results.

lame

Ignoring the core clock, the memory clock is beyond suspicious, it could be some kind of Nano-tier card, but there's no reason to underclock the memory that much.

Maybe pic related will die, Fury X was a inefficient housefire.

I feel good now with my GTX 1080 purchase.

That being said AMD only redemption lies in pricing. If its going to be around $300 it could easily beat GTX 1070 which is around 350-400

>45% higher compute performance than Fury X.

Source?????

Fury X and Vega have 4096 cores.

>HBM2 on a $300 GPU
doubt

>mfw i bought amd stock

videocardz.com/amd/radeon-instinct/radeon-instinct-mi25

And Vega has around 1600MHz clocks and a new architecture.

>that delicious denial
The constant disappointment of AMDrones is something I look forward to every year.

MI15 is Vega with 12.5 TFLOPs

>slightly better than a fury x
i don't think this is real

*MI25

>Hbm 2
>1000Mhz
Well maybe it could possibly be 700, but not really, and totally not 1Ghz

Journalistic integrity at its best, all it takes is to look at the enterprise Vega card tflops to see that this is bullshit.

This is either.
1. Some low power version ala Nano
2. ES, which makes more sense as neither core or mem clocks make sense

>barely above the 1070

This can't be real kek

>MI25
>feel on the amd compute meme numbers

If I fowler your logic, the rx 480 is faster than gtx 1080...

>my 7700k offers 10% more gaymen performance compared to 8 cores clocked 1GHz less!!!!!!!

>ITT: faggots arguing about a POSSIBLE score with no concrete tests
Sup Forums is always eager for fanboy wars. fuck the underage faggots in here.

>Card Status: Unreleased
>300 watt
>This card has not yet been released, specifications are subject to change without notice

Paper launch

3dmark.com/spy/1544741

>3dmark.com/spy/1544741

3dmark.com/spy/1544741

>3dmark.com/spy/1544741

3dmark.com/spy/1544741

>3dmark.com/spy/1544741

it's a 45% uplift over Fury X
OP shows 9%
OP is obviously bullshit

I thought we were finally dwindling down on the non-stop Ryzen shitposting now that it's by far the best choice with the 1600/1600x. I can't wait for Vega shitposting.
This makes no sense whatsoever. It has the same cores as the Fury X on a smaller node with only minimal gains. Either this is a super downclocked version or there really is a smaller Vega. The latter would be nice, I'd love them to offer a smaller Vega at the $300-$400 price range.

I'm pretty sure there will be 2 Vega cards out there. One to competive with GTX 1070 one to compete the 1080 (or maybe even the TI)

That being said, this looks like to compete against the GTX 1070, but seeing how the performance is relative the same, AMD only go for the cheaper pricing technique

>possible vega entry
ayyy, nvidiot shills hard at work.

The MI25 is not a dual GPU?

>One to competive with GTX 1070 one to compete the 1080
That would be rather pathetic, those cards have been on the market for almost a year now

>Just bought a 1080
Damn, what a relief.
Always bet on AMD to be a fucking failure.

anandtech.com/show/10905/amd-announces-radeon-instinct-deep-learning-2017/2

F U R Y X 2.0

Unless it has 32GB

>Just bought one of the worst price to performance cards out there.
Damn, what a retard.

this means absolutely nothing. If you're so eager for a fanboy war you can be patient for few more days and keep shitposting in the faggy muh Ryzen cores VS muh Intel Giggahurtz threads.

>C87F:C1
Lol, this thing is from fucking December

He said he bought a 1080, not a fury x

Yes that why I said pricing is where at it.

The 1070 is around $340-$400 on newegg. ($340 is a small version for ITX, the proper MSI/ASUS/Gigabyet/EVGA are all around $380)

If the Vega with the similar performance as the GTX 1070 goes sub $300 it may be worthwile, specially if you have a Freesync Monitor)

I just bought a GTX 1080 but no regrets, EU prices will be fucked over for the next 3-4 month after release

Why that mean nothing? How you can fake 3dmark ID?????

>Ryzen shitposting now that it's by far the best choice with the 1600/1600x
Yeah...no

>GTX 1070 goes sub $300 it may be worthwile

I don't see that happening, not with the leaked die sizes and the HBM2 memory

Have fun with your 4 cores at the end of this year when they're already getting starved right now.

How many times do I have to tell you to not reply to fucking bait threads? and sage every post if you're gonna reply.

>GPU could be underclocked
>The GPU BIOS could be tampered with to make it unidentifiable
>Could be an engineering sample
There is many possibilities and it is funny how impatient everyone here is to jump on the bandwagon of shilling with/against AMD.

What's worse is when it comes out it will perform 15-20% worse than it does now. Show me one case where modern AMD processors and GPUs haven't disappointed users in performance tests after release.

Nice memes kid

Have fun jerking off to Adoredtvs video's

there are likely going to be at least 4 variants of vega, 4gb/8gb cut/uncut mix and match.

at vegas absolute worst, it should be matching a 1080, especially if they focused almost entirely on alleviating the bottlenecks.

So Vega is a overclocked Fury X? That's nice.

HAHAHAHAHA, look this AMDrone damage control post

AMD co-created HBM2 and we are not sure how expensive its.

I remember when there was a type of Second gen Memory what was cheaper to produce (and was more efficient) than Gen 1 (Maybe DDR1-DDR2?)

>it's not real reeeeeeeeeeeeeee!

This is some pretty terrible falseflagging.

Also
>reading comments on anything but RWT

>gaymur forgetting the 140w TDP 6900k housefire
Top kek
>inb4 b-but MUH GAYMEZ

Then there's still he massive die size

Maybe we need moar wait™ or moar cores?

Is anyone dumb enough to believe this when you can see the clocks displayed in the same bench?

Look at this thread, Intelfags will latch to anything at this point.

kek

This guy is crazy
Everybody read the comments on news

Nah, fuck all the tubers, they all suck.

These are youtube-tier comments at best, how can you stomach that level of discourse?

>This is not real.

RX 480 has a Memory clock of 2000, he telling us that a Pro Duo has 1/3 of that.

I'm 90% sure this is Vega, seeing the Memory is at 700. Its most likely just a test version with a lower GPU and memory Clock

we know that passively the card hits what, 1525/1575, i forget what the pro card hits exactly, so the consumer version, that has a proper heatsink and 2-3 fans pushing air clcoks what, 20~% lower then passive?

I call bullshit on anything not clocking on par with the pro line till amd come out and tell us vega clock speeds.

>1200mhz
>eng sample.
>already proven to be on par with a 1080 with alpha drivers that aren't using its full potential.
> NVIDIA slaves in G try to bash it.

Good luck with that you useless niggers.

>know it's at least 12.5TFLOPs
>know shader count

It's not magic to calculate the clockspeed.
This isn't even up for discussion, consumer Vega may clock higher at stock or may not depending how much headroom they wanna leave or how much lower power they wanna go

*IF* Vega is on par with a 1080 = big disappointment

>ends up being a mobile chip due to the heavy underclocks on both clocks

I hate predicting prices but even so if its on pair with GTX 1080 at performance but cost $100 less, it may not be a dissapointment

But that being said, Nvidia just cut the prices of the GTX 1080 and call it a day

This is good. It means they have to sell the card for $350 max. I hope they'll do a Nano card this time around too.

And I don't want to do math, its some where in the 1500 range I listed.
At amds worst, its on par with a 1080, at where I think they will land, half way between a 1080 and 1080ti, and at best, playing ball with the 1080ti or titan Xp, really it all depends on where they land with alleviating bottlenecks.

Not really, but then again I want amd to fuck off from the consumer gpu market for a generation or two, just because of how much I hate nvidia owners and their 'fuck amd, now my nvidia gpu costs more' line of logic.

>I want amd to fuck off from the consumer gpu market for a generation or two

They already did that.

Wasn't there a leak with the same PCIID months ago in compubench at the same clockspeed and it was between a 1080 and 1080ti?

Hard to believe they didn't get clocks up at all in that time.

Read the rest of the post. If the card is on 1080 level with alpha drivers and showed off so early in its development. Then it should be at 1080ti performance level upon its release. If it was a finished product six months ago with all optimizations done the it would have been released by now.

Aka it's a 1080ti similar performer that will cost a bit less and that's a good thing for everyone.

But Doom is not an AMD game?

That's exactly what a Fiji card with 8GB memory would perform like at these clocks. It's not VEGA. It's not even a die shrink.