AMD Radeon Vega Frontier for 1,199 USD

videocardz.com/70299/amd-radeon-vega-frontier-graphics-card-available-for-preorder

Other urls found in this thread:

wccftech.com/amd-radeon-vega-nvidia-pascal-gp100-performance-benchmarks/
youtube.com/watch?v=qspdnAYiiug
twitter.com/NSFWRedditImage

WHAHAHAHAHAHAHAHAHAHAHAHAHAHAHAHA

>it's cheaper than gp104 Quadro
Noice.

>375W HOUSEFIRES

AYYMD IS FINISHED & BANKRUPT

AYYMDPOORFAGS CONFIRMED ON SUICIDE WATCH

>2 8-pins

holy shit

$1200 for a workstation Vega GPU?
Thats pretty big. Not too long ago a flagship workstation card could have been $3000.

I guess they want marketshare and want to push ROCm.

>Double 8-pin

>cheaper than quadro
huh

hahahahahahaha

Any AIB gp102 is also 2x8pin.

>workstation card vs geforce on pro benchmarks
>not compare with quadros
>375w
WEW LAD

I was expecting something to bridge the gap between 480 and 1080. Who are workstation GPU's actually for and why do they spend that much?
CAD engineers? I know maya, udk all only use the GPU for view ports but the final render is done on the CPU.

wccftech.com/amd-radeon-vega-nvidia-pascal-gp100-performance-benchmarks/

>If we take a look at some benchmarks posted by PCPerspective a while ago, the Quadro P5000 (a GTX 1080 equivalent card) manages to score 152.92 in Catia. Vega scores 135.75 in comparison which is near the P4000, that is a cut down GP104 SKU. In Solidworks test, the Quadro P5000 scores 168.11 versus 114.88 on Vega. The score is even lower than Quadro P2000 which is based on a cut down GP106 die.

Quadro P5000 literally destroys 375W card in performance, how embarassing

AMD doesn't even dare to compare to Quadro P6000 or Quadro GP100, which would be even more embarassing the margin of defeat

This shit is slower than Quadros and draw more power. Vega is Fury X 2.0 confirmed.

>draw more power
Did you measure it from the rail, Huang?

>amd gpu
>2x8 pins
>high tdp
>aio option

get a load of this shill

Since when power connectors mean anything? Fucking AIB 1080tis have 2x8pins.

375w is beyond PCI-E spec. Wut.

>BUTTMAD AYYMDPOORFAGS WITH NO PERFORMANCE DETECTED

AYYMD GOT CRUSHED, JUST DEAL WITH IT, FAGGOT

>TDP 300w-375w

...

>trying this hard
your bad at it m8

>Complaining about wattage
>AMD is claiming 50% improved IPC and 2x performance per watt

4096 cores @ 1600mhz is going to BTFO of nvidias flagship period

More than that, 375w card simply can't pass PCI-E certification.

AYYMD went full retard as usual and overclock the chip well out of spec because they can't compete in performance, just like the Fury X

Did you forget dustbuster?

>not the updated version with pascal

never forgetti

>HBC
so it isn't HBM2?

That's nothing nigger my card has two 8 pins plus a 6 pin

It is.
Nothing beats 5800 ultra.

16gb of hbm2

Fury X 2.0

Fury X was 275W.

>5800 ultra
youtube.com/watch?v=qspdnAYiiug

HOLY SHIT!!!!

>Fury X
>275W

This is my point, overclocked Fiji with HBMeme2 like we said 4 month ago...

Single GTX 590 is a 600w card on one PCI-e

That's it's rated TDP. Do you know the difference between TDP and power draw? If course not, fucking Sup Forumsedditor.
The original housefire.

But hes right, you're a shill

That's dual GPU board.
Where the fuck did 50% more transistors go?

>Where the fuck did 50% more transistors go?
I bet in the HBM die or lies

Only the devil knows this kind of suffering
>tfw the CPU is @ 4.9ghz with 1.554v too
You ever seen a system completely max a 1500w PSU?

I know you.
What fucking HBM die you fucking retard? Vega10 is ~500 mm^2 die you nignog. Its still 4096 ALU. Where did 50% more trasistors go?

>wccftech

People just don't know any better.

>comparing a workstation graphics card to a consumer graphics card
God, this board used to be filled with people who actually knew about technology. Now it's just children arguing over which video game box can get them the most fps in cow-a-dooty.

Pathetic.

It's videocardz

Running a fury full throttle on a synthetic furmark heater is kinda retarded.
When you thrash all your SPs it's going to get hella hotter because is has 4k of the damned things on a bigass die.

Comparing full load on raw geometry produces better real world results.

Plus: Holy fuck you can push 432 watts through a single die what the fuck

I blame AMD. They can't beat the QUADRO, so they bench Vega Frontier with the Titan X (pascal)

Even so the rated tdp is way beyond reasonable. If it don't deliver better performance than p5000 at least is a shit.
Also considering this whole arch, it seen fury 2.0. only time will tell if this is a reasonable going for gaming or pro work.
On paper fury was a beast but in practice it was just a heater.

>TDP 300w-375w
You do realise you are quoting Videocardz guesswork with no actual evidence, right?

Wait, isn't this the workstation card? Isn't this a great thing?
It's looking like AMD is doing a Vega for workstation and a Vega for consumers.

Yes, workstation Vega even comes first, even though launching gaymen ones is easier. Weird.

They should have compared it with p5000 because it will be on same league.

Anyway, Titan Pascal xp is actually almost the p6000, it's the same die, minus drivers.

That is no wonder. Perhaps Vega is too hot and perform below expected so another flop.

All Nvidia cards throttle heavily while running Furmark. 1080ti would consume 350w or more if it actually ran furmark properly

>2x8

You also can't go past 1.2v without modding since Maxlel.

Go shitpost somewhere else. Most cards sold with 2x8pins have TDP in the 250W range

Does it do cryptocoins?

Its on a smaller fab process it should run a lot cooler than a fury x
This is true but performance can be judged from it indirectly
Ayyy
Vega is a all new design it could be compared to Bulldozer to Zen

Any AIB 1080ti has 2x8. So?

>Wait, isn't this the workstation card?

Only when comparing it to gaming cards, and with a huge possibility that benches were skewed to favor it completely.

>Isn't this a great thing?

Not if you need a workstation tier card.

>It's looking like AMD is doing a Vega for workstation and a Vega for consumers.

It looks like AMD's marketing dept is just having another giggle m8.

It's best to just wait until release especially if the information is from wccftech.

Scientific simulations, for one. Engineering stress simulations. Anything that wants a lot of floating points crunched. Vega's party trick is perfect FP scaling with size. FP16 is twice as fast as FP32 is twice as fast as FP64.

If gaming chips are binned Vega, you need to bin enough chips to actually have supply. Low cost workstation card and late lower-end chips could mean Vega has very, very good yields.

It's also in new gaymacs pro.

That shit behind a display? God within a week the display will melt unless it's lead shielded.

Why?

Ryzen runs cooler than Intel chips, maybe AMD figured out power delivery.

You just add tiling and larger L2 for more efficient GPU. See Maxwell.

What?! 2x8 pins? Fuck that shit, I am not buying a new PSU for that

It does have very good yields similarly to Zen. The 14nm FinFET is fully fleshed out at this point. We will see a cut down version of vega on GDDR5 for RX 600 series of cards but that is more than a year out.
Work station cards also have to be very accurate. This is one of the reasons gaming cards are not ideal for them. Its also the same reason why in general work station cards perform much worse in gaming. Double precision performance is king in scientific calculations.

Vega10 is 1/16 FP64.

P6000 still owns the top gaming performance believe or not.

I think Titan Xp is better. Both are full gp102.

P6000
>3840 CUDA cores, 12 TFLOP

Titan X Pascal
>3584 CUDA cores, ?? TFLOP

I said Titan Xp, which is 3840 ALU one.

Titan X (pascal) =/= Titan Xp = P6000

...

Fucking nVidia naming schemes

So if this workstation card is $1200, does that mean the regular Vega 64 will be more like $599?

I just sold both my 480's to coinfags on eBay and made a decent profit. Limping along with an old 960 for now. Should I get a 1080 or wait for Vega? I don't have any brand loyalty, just want to make the best choice. If I have to use this 960 for more than a month I'll probably neck myself though.

Really doesn't matter, 150w each from the two 8 pins + 75w from the slot = 375w. Not like the motherboard knows the difference.

Just makes buying PSUs more difficult.

Agree

>If I have to use this 960 for more than a month I'll probably neck myself though.

RX Vega is at least 6 weeks away.

It does matter when you need PCI-E certification.

It doesn't matter at all, two 8 pins means it's still only drawing 75w from the board. Just makes it harder to find a PSU.

>--
Pathetic. even they know how embarrassing it is.

>Need
That's a funny word

The new iMac will probably be the cheapest way to get a Vega machine

AMD already has tiling, nvidia's is just better.

Current GCN tiles are too big.

I think they're supposed to be like a 1/3 of the scene or something, or divide the scene in 3s (so 9 tiles) or something. I vaguely recall 1/3 being tossed around.

It's 1/4 of the scene.

DOA

>need PCI-E certification

>videocurryz
Still don't know better

And another 6 months before you can actually get your hands on one