AMD DID IT AGAIN

File: 1478541707353.jpg (11KB, 251x242px) Image search: [Google]

11KB, 251x242px

>>61098602
>it's another AMD shill thread

so 1070 performance level shitposter was right? fucking A we will never hear the end of this

>>61098602
>From what we’ve seen, that concern may be misplaced. It appears to be plenty fast and, at least for the settings and the games we played, indistinguishable from the competition. Our original estimates after seeing Radeon Vega Frontier Edition with Sniper Elite 4 at Computex still hold: The cards appears to be faster than Nvidia’s GTX 1080 and close to that of a GTX 1080 Ti card.

File: 1498197847904.jpg (58KB, 550x550px) Image search: [Google]

58KB, 550x550px

>>61098602
I will say it again. VEGA IS FURY X 2.0 tier, I hate the AMD overhype

>>61098681
were they aiming at the skybox like last time?

>>61098681
>frontier edition, which is confirmed to not be for games, and in fact, supposed to be worse for games than the consumer release, is almost taking out 1080ti

Holy fucking shit, I didn't expect this

>>61098722
Like that MSI rep said about the power draw .. "It depends on the performance".

>>61098722
It's a 1500 card I would hope so

>>61098699
>I don't have reading comprehension, the post

Are you an idiot, or just pretending to be stupid? Vega frontier edition is beating 1080 in that benchmark, and pretty handily, at that.

>>61098681
1080ti BTFO by $399 Vega confirmed

>>61098747
Workstation card. That's par for the course. Consumer Vega supposedly has an "excellent" price, and be better for gaymes.

If this thing really is priced that good, Nvidia really is going to be in a pickle

>>61098602
AMD/Ati always scored higher on synthetic benches, but worse in games.

The card will be barely on par with the 1080.

>>61098800
Great job not reading the article or the thread.

>>61098781
>the miners will devour this cards production runs

>>61098800
That was before, when they couldn't handle Goyworks' hard-on for Tessalation

repeat after me:

1070 level performance

>>61098602
>pre release Vega is radically faster, by an immense margin, than the Titan X Pascal
>we better find a way to spin this as a negative

Lol Jesus the press is terrrible.

>>61098841
1080ti level performance
(for less than half the price)®

>>61098851
announcing reports is a bannable offense. want to push your luck kid?

>>61098861
delusional. they can't even release it on time and pumping it with power to meet expectations.

>>61098750
I did not imply the opposite, I say Vega is not efficient nor will beat the 1080ti, It a Fury X 2.0 tier

>>61098823
Isn't hbm bad for miners?

>>61098868
Anon never said he was reporting anyone, Anon.

Just a general reminder to give janitorial staff / MODS a heads up to aid them in their work.

>>61098877
>muh TDP argument
Why do goyim always default to this, when they get scared?

>>61098893
No idea

Seeing as different cryptocoins like different setups, who knows until it's tested.

>>61098893
They're probably too retarded to even know what that means. They'll fly off the shelves because the AMD's most recent card.

>>61098906
What is Fermi? What is gaymen laptops?

>>61098906
because I have a fermi right now and don't want another space heater gpu when I upgrade

File: 1498509832864.jpg (57KB, 711x533px) Image search: [Google]

57KB, 711x533px

>>61098877
If the end result meets the 1080 ti, and has a lower price, I think it's worth it. Also, keep in mind the 1080 ti comes pretty close to 300 watts anyway. And doesn't have a hardware scheduler.

>>61098781
this

>>61098923
Fermi wasnt just high, it was a disaster.
>>61098928
I'm so sorry

if it's 1080ti levels explain why they have been so quiet about it's development and that disastrous tearing prey demo

>>61098823
stores where i am will have limits at least for the first couple weeks or something to prevent coinfags sucking up all the stock

>>61098973
dont know why they even care to hold anything back

they must know coinfags will buy every amd card they can lay hands on so no worries about sales

>>61098973
well by the prey demo the fucking thing should be 1060 level so it makes no sense

>>61098951
>460 and 460ti
>disasters

>>61099052
>460ti
560ti*

>>61098763
Fucking this, people don't understand this is exactly what AMD did with Ryzen. They flat out lied to Intel about it's capabilities, and led them to believe it was going to have 10-15% worse IPC than it ended up having.

>>61098893
>Isn't hbm bad for miners?

It's bad because the only card on the market with HBM is limited to 4gb, which is bad for the miners.

With 16gb hbm2 it won't be a problem.

Never fall for the AMD hype.
Vega will be 20% slower than the 1080ti.
Faster in AMD titles
At least $500

1 year later when drivers finally mature, it will be on par with the 1080ti.
By that time, the 1180 will be out and AMD will be 2 years behind again.

This shit happens every fucking time.

>>61099698
You bring up an interesting point there mr.shitposter, although not the one you'd think. If I'm spending $500+ on a GPU, I'm sure as hell going to want it to stay relevant for more than a few years before it becomes obsolete. Nvidia is practically known for this shit. I'd expect a GPU that expensive to last for 5-6 years, at least.

>>61099814
You're supposed to sell your GPU every year and buy a new one, that way GPU's only cost about $100 each.

>>61098615

It's benchmarks fagget, numbers don't shill.

With computer components, we get to measure which things are better at which tasks, and in this instance Vega is better than Titan XP.

>>61100191
>and in this instance Vega is better than Titan XP

Did you read the OP post? Vega is slower.

>>61100208
>Did you read the OP post? Vega is slower
Did you read the OP link? Vega is faster

>>61100315
http://www.pcworld.com/article/3202659/components-graphics/hands-on-amds-radeon-vega-frontier-edition-vs-nvidia-titan-xp.html

>but to be fair, it’s comparing an ostensibly consumer-focused Titan Xp card to the pro-level Frontier Edition. Nvidia’s Titan Xp, for example, officially works only with consumer drivers, while the company’s Quadro workstation products offer a lot more performance

http://wccftech.com/amd-radeon-vega-frontier-edition-preview/

>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 137.68 in comparison which is near the P4000, that is a cut down GP104 SKU. In Solidworks test, the Quadro P5000 scores 168.11 versus 111.13 on Vega. The score is even lower than Quadro P2000 which is based on a cut down GP106 die

>>61100315
They cherrypicking the benchmark and refuse the bench gaymes...

>>61100400
titan xp is the giant pascal chip right?
how do you come to a conclusion that it will be slower in games than 1080ti which is slower than titan xp?

Personally i'm okay with it if it's 10% slower than 1080ti but costs 600 instead of 750.

>>61100458
I've seen several 1080ti sell for $650 though.
The GPU market is kinda fucked right now.

>>61100458
>1080ti which is slower than titan xp?

Because the Quadro P5000 (GTX 1080 with pro drivers) is faster and the Titan Xp/1080ti don't have pro drivers like the Quadros.

They did the same shit with Ryzen...

>>61100509
isn't FE supposed to have pro drivers? I can't remember what Raja said on analyst day.

>>61100578
It is. Read the review

http://www.pcworld.com/article/3202659/components-graphics/hands-on-amds-radeon-vega-frontier-edition-vs-nvidia-titan-xp.html

>>61098699
...The Fury X wasn't a terrible purchase if you got a good price on it.

>>61100599
>ran one of the first air-cooled production-level Radeon Vega Frontier Edition cards AMD has built

air-cooled version.

>>61100509
>They did the same shit with Ryzen...
yeah, they compared 8 core to 8 core, how terrible

now they compared $1200 GPU to $1200 GPU

>>61100112
Lol. What kind of math is this bull shit?
People don't believe this do they?

>>61098602

Fell for the hype with the RX480, not tricking me this time.

>>61100638
>you are supposed to sell your car every time ashtray fills that way every new car would cost you only $10K

>$1200 300W
>Beats a 1080ti

Wow, awesome. Sign me up.

>>61100678
lol it's gonna b $299

source: dad works for amd

>>61100509
So you would rather they compared a $1200 card to a $5000 card to be (((fair)))? And here I was thinking shills were a myth.

>>61100668
The RX480 is legitimately a good card when it's not priced out by cuckminers. Did you expect it to beat 1080s or something?

>>61099113
Then that just goes to show that Intel's holding back and not giving the consumer what they really could offer.

>>61100731
people expected better perf/watt. since they hype it so much

>>61100754
RX480 is 95w TDP card in oem/embedded form (as the E9550)

but after that market scoops up all the top binning chips, gaymurs and coinfags get what's left with the voltage cranked to harvest as many dies as possible or whatever

most cards could get away with a lot of undervolting because of this

File: 1461262274864.jpg (41KB, 688x456px) Image search: [Google]

41KB, 688x456px

>>61098615

>>61099599
4GB it's still plenty for Ethereum mining, nothing else needs so much memory
Ethereum might drop GPU mining by the time 4 GB isn't enough

>>61099113
But it hasn't really worked with Nvidia, they aren't sitting on their asses pretending they have secured complete dominance
Hell, they don't even have complete dominance

>>61098699
>>61098602
The Vega chip is far cheaper than the quadro GPUs

>>61100874
Except even Volta is rumoured to be without async computing?

>>61100898
That hasn't stopped them from putting out a giant 800~ mm2 die with meme shit like trashy deep learning accelerators, they know they need to R&D and have been putting products out
Intel didn't even try anything until Ryzen was weeks away from launching

ctrl+f p6000
0 results
Cherry picked AMD shills at it again.

>>61100898
Who needs Async with their hardware being so efficient?

Dx12 and vulkan only remove AMDs inherent issues with several overhead and shader usage.

>>61100707
And I‘m rome‘s pope

>>61101204
>p6000
you mean to compare $6000 GPU to $1200 GPU
that would be fair!!

>>61101293
can you please be less of a heretic

k thank

>>61101301
I mean to compare the fastest Nvidia have to the fastest AMD have?
Yes.,

>>61101305
F off nigger

File: 1497249016568.png (34KB, 769x733px) Image search: [Google]

34KB, 769x733px

>>61101312
>Comparing a $1200 GPU to a $6000 GPU
>Openly admitting you're that dumb

>>61101312
they did compare it to titan xp, it is the fastest in that price bracket
stop pretending prices do not exist
it's not AMD fault nvidia pulled back everything good about titans and made them gaymer only cards

>>61101335
A fucking tripfag

>>61101335
>>61101340
>It's NVIDIA's fault AMD doesn't have good hardware
Titan Xp is worse than P6000
And I'm ignoring prices because it's about what is faster.
It's a benchmark topic, not a poorfag topic.

>>61101359
>Titan Xp is worse than P6000
aren't they the same gpu?

but then quadro has ecc ram and the workstation drivers tuned for the pro software instead of gaymes

>>61101359
>It's a benchmark topic, not a poorfag topic.
no anon, it's what people will buy topic

>>61101382
Vega FE has pro drivers
P6000 has pro drivers

Vega FE beats Titan XP which does not have pro drivers, but loses to P6000 which does.

Presumably RX Vega without pro drivers will score less than a Titan XP, so Vega is probably lower performance than a 1080ti

>>61101397
release date is today right? will know in a few hours

>>61101246
>who needs technological advancement?
>just buy our new rebranded product, goyim!

>>61101419
Today?
Perfect, I can now play Starcraft HD without lagging.

>>61101472
The NDA end today!

>>61098602
Let's just ignore one card has professional drivers which boosts performance massively in these applications, and that a pro Polaris 10 has the same performance as Vega

So why are they comparing pro vs normal drivers?

File: meanwhile at amd.jpg (46KB, 500x486px) Image search: [Google]

46KB, 500x486px

>>61101944
>pro Polaris 10 has the same performance as Vega
hmmmm, I think you are onto something

>>61102007
https://forums.anandtech.com/threads/vega-navi-rumors.2486940/page-171#post-38946534

There's Polaris 10 WS/Pro comparisons near the bottom of the page.

>>61102033
So what the fuck is FE Vega for?

>>61102091
Early adopter tax and jack of all trades, good in gaming and pro but not best in either, Pro Vega and RX Vega will be shit at pro and gaming respectively.
And across the board faster in fp16 than Nvidia.

That's why it costs so much.

>>61098602
i think this article just sealed my assumptions i did on beyond3d

those vegas arent pure compute or pure gaming they are basicly the titan line of amd prosumers cards
it is indeed strange that amd is sandbagging so hard on vega
i expect that a variation of MI25 will be the true compute card and the MI16 will be the top of the line gaming card(given the tdp and perf as well)

>>61100885
both p2000 and p4000 are cheaper

>>61102390
My bad, was thinking of the other quadros, Vega is DOA for some production applications simply due to CUDA, but I wonder how the newer WX cards stack up against those.

>>61102091
Titan (the original, "prosumer" one) from AMD.

>>61098602

I'm really not getting this Vega FE. AMD keeps advertising it as a workstation card, but the more results pop up, the more it looks like AMD's take on the Titan than a true workstation card. Even a WX7100 scores 100+ in Catia, and given the chip it uses, you'd expect the gap to be larger.

This is some 4D Chess, FE is there to obsolete the Titan, forcing Nvidia to put out a prosumer GPU like the Kepler Titan again in its lineup so people don't buy their quadros.

>>61102584
It's literally AMD's Titan (the original kekler one).

File: weak_fear_the_strong.jpg (170KB, 850x851px) Image search: [Google]

170KB, 850x851px

>>61098602

>>61101359
>a $200k 8 socket E7 system performs worse than a $12M mainframe in certain workloads Intel BTFO of the server/enterprise market
Does what I just saud make any sense?

>>61102584
Think less "Pro CAD" and more "I'm paid to fuck around in UE4 or Unity" kind of workstation

Like everyone's saying, it's a kind of in-between card like the original Titan was. Being passable at the pro shit is handy, while being more aimed at heavy compute and doing serious shit using game engines.

I wonder if the RX cards will still give us FP16 and INT8 support, because they look interesting from a compute perspective (especially outside of meme learning shit, stuff like image processing could be done really fucking fast when you can drop precision for 2x or 4x speedup)

It's a 1200$ workstation card just for the air cooled one and although they say they expect high support for games, it's not gonna matter because it's designed to crunch numbers in a workstation environment not 4k gaymin.

>>61102695
weren't kepler titans really popular before nvidia shut down that store for higher margins to be cool like intel for them margins?
I think AMD got a shot here, market is a bit starved for such a card
But hey, now is perfect time to release new arch for them, they will sell everything foundry will pump out either way.
Can't wait for q3/4 financials.

>>61101359
>Caring about blind brand loyalty instead of market segments
Post a timestamped picture of your P6000 Quadro, if you're so ghetto rich.

>>61100615
>if you got a good price on it.
Ding ding

That's the problem, it originally retailed for $650 to compete with the 980 Ti. At that price it just couldn't compete.
>Less performance
>Less VRAM
>Less power efficient
>Issues with pump noise

I own a Fury X, and I didn't get it for $300 I wouldn't have gotten it.

>>61102781
Yes, most of the bitching for later Titans post Kepler were gimping of FP64 functionality.