Annnd the RX Vega is a huge disappointment, according to Guru3D

[RustyKats]

FuryX is 4096bit HBM1. VEGA is only 2048bit. So there is probably little to no change at the GPU design. NCU? More like GCN2017

Jesus dude you sound like a paid shill to crap on AMD based on leaked ES samples running low clocks without final drivers or bios... quite sad man.

Did you forget Polaris ES was leaked at 800mhz? And people like you were claiming it means AMD would fail...

And look where RX 480 are, matching or besting Pascal 1060.

 

[HD64G]

Not possible that the full Vega product could be so low in performance to just surpass Fury-X by few %. Not simply my 5 cents...

 

[Ebo]

Has anyone even considered that this is comming out, simply because AMD has been so tight lipped about what Vega can do ?.

Nobody knows yet what the new arkitecture brings to the table, and sites still has to post what Trump calls "fake news" just to keep the ball rolling until the card is actually released.


For me all that counts is how Vega preforms in 4K, nothing else matters, simply because I have a 55" 4K tv I can use as a monitor if Vega does well. then my wife can see her shows on the old 32" 1080 tv .


Im considering Vega because of new arkitecture and new ram tech and can it trade blows with GFX1080/1080TI Im a happy camper. If it cant surpass those cards Im still considering it, just for the fact that its new tech.

 

[ShurikN]

Ok so let me get this straight.

Same amount of SP on both cards. Fiji has 1050Mhz, Vega (this version) 1200Mhz. That's a 14.3% increase.
Double the mem bandwidth on the Vega.
Not to mention lower process, shitload of new technology, and the card is only 9.7% faster in this particular test... cmon ppl you can't seriously believe this bullshit.

I mean, if AMD literally just took Fiji, put it on 14nm and stay with HBM1, it would perform better than this card.

This card also clocks at 1.2GHz, even Polaris 10 can easily reach 1.4Ghz with adequate cooling and good power delivery.

 

[the54thvoid]

Lol.

And believe here actually believe this Guru 3D rumour is a real indication of performance.

It's another day at Gullible HQ.

 

[IceScreamer]

From reddit, this is the 3Dmark comparision of the RX480 and what people fell for is the Vega. http://www.3dmark.com/compare/spy/232689/spy/1544741 so yeah, take it as you will. My guess is this could be... I honestly don't know.

 

[Vya Domus]

I can hardly believe a new architectural design meant to fix all the flaws that Fiji had on a new node at presumably higher clocks can barley outperform the 2 year old Fury X.

So this is either the lower end Vega chip ( still somewhat unlikely ) or this is complete bullshit. Lets ignore this one , shall we ?

 

[uuuaaaaaa]

The Radeon Instinct MI 25 (server part) has 12.5TFlops which means 1525MHz, I imagine that the desktop part would be clocked at least the same, most likely higher than 1525MHz. Also the Vega architecture appears to be quite different from polaris, which means it cannot be directly extrapolated from other AMD cards. I've seen these results earlier, and I do not believe that Vega will be this bad. Also those 700MHz put the HBM2 slower that the HBM1 on the Fury X

 

[RejZoR]

People seriously believe this is RX Vega, the top model? Even Fury X made on old manufacturing process and constantly thrashed by everyone how "bad" it is, didn't miss the mark by this much. And people think this is a real thing. God help us all.

 

[Estaric]

I know I haven't been apart of the "PC gaming" world nearly as long as others but do we really want this to be true? I have friends that when I even mention the fact that I run an AMD card they get all mad and confused about why I purposely waste my money on crap products. I have ran 3 AMD cards 280x, 290x and Fury and they are still gaming very well. I'm just trying to see what anybody gains from if AMD fails, to me it would just give Nvidia a reason to up prices again. And none of us want that.

 

[Xuper]

The best logic that AMD still would not release ANY performance related information on VEGA is it is a huge disappointment. They probably has been trying any possible way to make it look less pathetic. But software optimization can only do so much when the hardware design is not improved. The HBM2 version will probably be between 1070 and 1080 level performance.

So much for their "Poor Volta"

I am switching to team green soon. AMD's GPU, well, belongs to the past.

Good GPU design requires resources. With ghe majority of AMD worked on RyZen RTG had neither the brain power nor the money to come up with something spectacular. People use RyZen as example to show AMD can pull off something great, at the same time they ignored the fact that AMD simply don't have the R&D budget to tackle both CPU and GPU design. Pascal was great, because Nvidia has all its gigantic R&D devoted to it, while they happen to have real talented electronic engineers. It would be almost impossible for AMD to give same amount of performance improvements with a tiny GPU R&D budget. For AMD, only competing at low end is the best option. Same for rebrand as much as possible to save on R&D. I feel like RTG is going to get smaller and smaller and eventually AMD is only going to focus on APU level performance tier.

I would wait for Volta, and I am waiting for Volta.

So much for your unbelievable childish behavior!

People seriously believe this is RX Vega, the top model? Even Fury X made on old manufacturing process and constantly thrashed by everyone how "bad" it is, didn't miss the mark by this much. And people think this is a real thing. God help us all.

lol , People can't control their emotion!

 

[oxidized]

Nothing is out of reach when you have zero facts... Ryzen was immensely tight lipped and performs admirably. I see zero reason to count the chickens before they hatch and look forward to the cards.

Here is to hoping. If not guess I'm going 1080tis under water

Tight lipped, really?? I can't even remember the countlews leaks and rumors coming out, where ryzen had insane ipc, kaby lake levels gaming, and huge overclockability.

Anyway i agree with the facts that we should wait, as we know nothing really tangible, surely this disappoints if true, and kills hopes a bit, to me Vega has to be at the very least on 1080ti levels, otherwise is a failure at least from an architectural point of view, this said, they might keep prices very low, but won't change the reality of the design, same thing as polaris, good-ish card for the price, but design is a bit poor, especially the refresh.

Still decent hopes for Vega tho

 

[RejZoR]

People seem to have already forgotten all the piss they took at AMD when first "mysterious" benchmarks of Ryzen clocked at 1.8 GHz surfaced. It'll be absolute shit they said, AMD can't clock them higher than that they said. And here we are with 4GHz Ryzen processors that are great in every possible metric. Let that be a reminder for the future...

 

[Assimilator]

Ryzen was immensely tight lipped...

No it wasn't. That 5GHz CanardPC leak was merely the tip of the iceberg.

Past experience has shown that when AMD has the performance to match the hype, they leak benchmarks. When they don't have performance, they leak nothing. And Vega has very definitely been on the quiet side.

I'm not enough of a fanboy to write off Vega until I see actual third-party benchmarks, but deafening silence is rarely good.

 

[FordGT90Concept]

It is an undetected AMD card, is has 8GB of graphics memory and the strongest indicator of it all, the memory runs 700 MHz. In case you don’t remember, HBM memory on the Radeon R9 Fury ran at 500 MHz."

I thought nominal clockspeed for HBM2 was 1000 MHz. I think they have the GPU and memory underclocked for benchmarking. CompuBench scores tell a similar story with 687F:C1 at similar clocks to RX 480 scores well below:
https://compubench.com/compare.jsp?benchmark=compu20d&did1=45270653&os1=Windows&api1=cl&hwtype1=dGPU&hwname1=AMD+687F:C1&D2=AMD+Radeon+(TM)+RX+480

The really interesting thing is how poorly it scores on Catmull-Clark Subdivision Level 3 but it blows RX 480 out of the water in Catmull-Clark Subdivision Level 5.

 

[RejZoR]

They were super quiet for Ryzen. The stuff you've seen were just bunch of clueless leaks and speculations, it was basically right before the release when AMD started bragging about matching 6900k. And they delivered.

If you recall RX 480 on the other hand, it was never a secret that it'll be a mid end part. It was again just bunch of wild speculations how AMD will pull insane efficiency. I was like, dudes, they are aiming affordability, you seriously expect them dumping millions of R&D into a mid end part? Really? It was cheap and just enough efficient to match the price they were targeting. It's why I found all the whining over slightly higher consumption even more funny.

 

[m1dg3t]

I'll just plop this here

Seriously LoLing at all the peeps claiming AMD is doing all this hyping, AMD has been lips sealed until the last minute with Zen and now Vega. The community is/has losing/lost it's shit and is making up all kinds...

MI25 is 12.5 tflops & ~1500Mhz yet the consumer version clocks lower and crunches less? HaHaHaHaHaHaHaHaHaHaHaHa

 

[ERazer]

What does ppl expecting from vega? same performance as 1080?

 

[RejZoR]

@m1dg3t
Not to mention MI25 is PASSIVELY cooled chip. People don't seem to realize how power efficient something has to be to deliver such compute power and be cooled passively. There are some differences that also make up the price, but still, it's telling.

 

[Vayra86]

Just putting this thought out there:

If Vega was a highly improved architecture, why rebrand the RX480 on an inferior arch and run with it for the coming year.
If Vega would clock considerably higher, why doesn't the RX580 clock well and why is its efficiency so bad.

If Vega is not a highly improved architecture, all they can do is bake a Fury on a smaller node. Which is exactly what it's gonna be. Let's not fool ourselves.

What we see in this chart here is likely the bottom of the stack, and the top of the stack will land at GTX 1080 perf. Best case.

Mark my words.

I wouldn't have bought a 1080 last month if I knew Vega was going to kill everything

Economically, the release of Vega serves only to portray the image of AMD still playing with the big guys. I wouldn't be surprised if it gets priced stupidly high for its performance, just like Fury X at the time. Its a massively complicated chip if its a wide design like Fury, which it almost has to be, otherwise we would've had Polaris in a larger package, and we would've had it last year.

Honestly, there's no wiggle room here.

 

[biffzinker]

What does ppl expecting from vega?

I know I wouldn't mind getting 1070 performance with the price around $275-300

 

[alucasa]

Just putting this thought out there:

If Vega was a highly improved architecture, why rebrand the RX480 on an inferior arch and run with it for the coming year.
If Vega would clock considerably higher, why doesn't the RX580 clock well and why is its efficiency so bad.

If Vega is not a highly improved architecture, all they can do is bake a Fury on a smaller node. Which is exactly what it's gonna be. Let's not fool ourselves.

What we see in this chart here is likely the bottom of the stack, and the top of the stack will land at GTX 1080 perf. Best case.

Mark my words.

I wouldn't have bought a 1080 last month if I knew Vega was going to kill everything

I really wanna do your avatar on a lot of people in this thread.

 

[Vayra86]

I really wanna do your avatar on a lot of people in this thread.

Perhaps I should do this

 

[Vya Domus]

Just putting this thought out there:

If Vega was a highly improved architecture, why rebrand the RX480 on an inferior arch and run with it for the coming year.
If Vega would clock considerably higher, why doesn't the RX580 clock well and why is its efficiency so bad.

If Vega is not a highly improved architecture, all they can do is bake a Fury on a smaller node. Which is exactly what it's gonna be. Let's not fool ourselves.

What we see in this chart here is likely the bottom of the stack, and the top of the stack will land at GTX 1080 perf. Best case.

Mark my words.

I wouldn't have bought a 1080 last month if I knew Vega was going to kill everything

Economically, the release of Vega serves only to portray the image of AMD still playing with the big guys. I wouldn't be surprised if it gets priced stupidly high for its performance, just like Fury X at the time. Its a massively complicated chip if its a wide design like Fury, which it almost has to be, otherwise we would've had Polaris in a larger package, and we would've had it last year.

Honestly, there's no wiggle room here.

Sorry but your seriously going off track here.

First of all AMD has gone out and released official information about architectural changes. Vega is not Polaris , I'm not saying it's miles ahead of Polaris but they have been very clear about it. I really don't see any company such as AMD going out of their way and talk about random gibberish such as blast processing in this day and age.

Secondly , no one said it will clock much higher than Polaris , but chances are that there is an uplift in frequency.

And last of all , if it's nothing more than a Fury X on an improved node , why would it take so much god damn time to release it ?

I think you have been shot point blank by the hype gun and can't filter out information by yourself.

 

[eidairaman1]

Prejudice is so common here now. This is just another fud thread.

Best thing to do, wait till users here test it for themselves.

Lol.



And believe here actually believe this Guru 3D rumour is a real indication of performance.


It's another day at Gullible HQ.

+1