Vega/Navi Rumors (Updated)

[guskline]

I'm with Mopetar. AMD/Gloflo was probably already working on a new revision. Their launch date comes up and they decide to launch with the current revision rather than delay for several months. It would make sense given the power fix they implemented. I find it hard to believe that the engineers missed something related to the power delivery. More likely they designed/programmed the first release to target values that didn't hold across all silicon. Either way we need more data to determine if this is truly a new rev or just a really good piece of silicon. My gut feeling is that any process issues will be nailed down by the time Vega releases.

I hope you are correct.

Vega needs to hit the ground running solid and closer to the high end Nvidia GPUs.

 

[William Gaatjes]

Tile based rasterization, that reminds of the powerVR/ ST kyro II i still have laying around.
But that requires a lot of local storage cache for all those rasterization units, yes ?

 

[PhonakV30]

XFX RX 480 GTR on air: 1288 MHz@ 1.05v around 85-90W in Heaven benchmark, 1475 MHz@1.185v - 133W.
Water Cooled RX 480: 1266 MHz@1.05v around 105W in heaven benchmark, 1470 MHz@1.185v - 150W, and unstable/crashes.

Gigabyte G1 from the same reviewer:
125W at stock clocks. And 180-190W when OC'ed.

Something is just not right here, guys...

I do think that XFX RX 480 must use at least new revision of the process, or be made on Samsung process. Then it is logical.

I have derailed the thread a bit, while staying a bit on the edges of the topic(Vega achieving high clocks), sorry for this.

anyone can check the First version of XFX Rx 480 then Compare it to RX 480 GTR.they should have different numbers like Revision.maybe from Bios or something like that

 

[Glo.]

But that requires a lot of local storage cache for all those rasterization units, yes ?

Yes.

Using Tile Based rendering would require redesigning the caches, also.

 

[Snarf Snarf]

Vega needs to hit the ground running solid and closer to the high end Nvidia GPUs.

I'd take that a step further, Vega needs to hit the ground going 100 mph lol. 6 months late to the party, if it doesn't definitively beat the GTX 1080 with enough headroom to make a push at GP102 AMD is going back to sub 20% market share. They can't keep selling big dies at bargain prices to get market share up, they'll bleed out faster than they already are.

 

[William Gaatjes]

Yes.

Using Tile Based rendering would require redesigning the caches, also.

And preferably, the caches are independent. Each raster unit would need to have its own dedicated cache. That way, less stalls, the better.

 

[Mopetar]

I'd take that a step further, Vega needs to hit the ground going 100 mph lol. 6 months late to the party, if it doesn't definitively beat the GTX 1080 with enough headroom to make a push at GP102 AMD is going back to sub 20% market share. They can't keep selling big dies at bargain prices to get market share up, they'll bleed out faster than they already are.

They can't sell big dies either if the process isn't mature enough to support them without being power hungry messes that need impeccable cooling systems. Better to release a flop that would only lose mindshare. Lacking a product, fans can at least speculate how good it will be in the future. If you deliver a turd to their doorstep, they might stop being fans.

 

[Snarf Snarf]

They can't sell big dies either if the process isn't mature enough to support them without being power hungry messes that need impeccable cooling systems. Better to release a flop that would only lose mindshare. Lacking a product, fans can at least speculate how good it will be in the future. If you deliver a turd to their doorstep, they might stop being fans.

I completely agree, it's better to delay than to deliver a turd. Vega release for me is the earliest I'll be looking at for upgrading over my 980ti. If 1080ti and Vega aren't at least 50% faster and under $700 I'll be sitting out this generation until Navi and Volta.

 

[geoxile]

There was that one poster on OCN who claimed AMD royally messed up Polaris and it wasn't even hitting 1000mhz. IIRC he claimed Polaris would be delayed until late 2016 to early 2017, so maybe he was partially right and there is a new revision.

Edit: I think he might've said delayed till October, in which case he would be right on the money.

 

[jpiniero]

I completely agree, it's better to delay than to deliver a turd. Vega release for me is the earliest I'll be looking at for upgrading over my 980ti. If 1080ti and Vega aren't at least 50% faster and under $700 I'll be sitting out this generation until Navi and Volta.

1080 Ti being 50% faster is possible... Under $700 no. One of the earlier leaks had it being $899 but if it's anything like 1080 you won't be able to buy anything near that. Vega's price obviously depends on its performance vs GP104/2 but I wouldn't expect much of a discount given how expensive HBM2 is.

 

[ConsoleLover]

When Vega ships, considering it would be over 6 months late to the party, possibly even more, the highest end chip needs to actually beat Nvidia Titan x Pascal version or at the very least be equal to it, but be able to lover prices and beat it on value. Considering the Titan X is $1200 I don't think it would be hard to beat it in value, but still, the new GPU's need to be really fast.

They can't have their top end chip only compete with the potential 1080ti, when Nvidia is likely to come out with revised versions of the cards and offer 10% or 15% more performance whether due to clocks or simply one upping the cards. Sure its probably going to be March/April before they come out with revisions, but considering Vega is only mid Q1, it doesn't give it enough time to shine, even if its on par with titan x.

They have to offer significant performance increase and beat Nvidia this time around.

 

[nathanddrews]

So... any rumors about Vega or Navi worth knowing?

 

[Snarf Snarf]

1080 Ti being 50% faster is possible... Under $700 no. One of the earlier leaks had it being $899 but if it's anything like 1080 you won't be able to buy anything near that. Vega's price obviously depends on its performance vs GP104/2 but I wouldn't expect much of a discount given how expensive HBM2 is.

I would expect $899 without any pressure from Vega. If there's only founders available for months and nothing but inflated prices on aftermarket cards at launch Nvidia is going to pull off another smash and grab. I badly want prices to stay sane for this hobby but things really are getting out of hand if a generational upgrade goes from $699 980ti to $899 1080ti. The real dream is big Vega ~1080ti at $699 like Fury at launch, cut down at $499 small Vega at $399 and cut down to overlap Polaris or to drop its price to actual $250 while small Vega cut down is $299.

 

[Dave2150]

Hoping to see some leaks of the Vega lineup before Christmas.... currently on a 1070 but eager for an equivalent/faster AMD card to use with my Freesync monitor! (XL2730Z).

 

[godihatework]

So... any rumors about Vega or Navi worth knowing?

Nope

Sent from my Nexus 5X using Tapatalk

 

[Det0x]

http://wccftech.com/amd-dracrays-vega-10/

WCC says:

"We’ve just learned about AMD’s craziest graphics product to date. A ten thousand dollar behemoth powered by Vega 10 & 1TB+ of memory & we’re going to share all the details with you today, exclusively.

This next generation mini super-computing fire-breathing graphics monster is code named “Dracarys”. The name, incidentally, translates to “dragon fire” in High Valyrian, a language in Game Of Thrones lore. As of today, we’re told, that early Dracarys prototypes are up and running and ready for primetime. In fact, our sources tell us that we’ll get a sneak peak at this Vega 10 powered graphics board as early as next month at Supercomputing 2016.

Vega 10, The Pulsating Heart Of AMD’s New Fire Breathing Monster
AMD’s next generation Vega 10 GPU will power this unique board which will also be equipped with an excess of one terabyte of directly accessible on-board graphics memory. Represented in 16GB, interposer linked, 2.5D stacked, second generation high bandwidth memory in addition to an on-board solid state device. So, in many ways, Dracarys is quite similar to AMD’s recently announced Radeon Pro SSG. Which features a Fiji GPU as well as an integrated SSD storage option, expandable via an M.2 slot.

However, the storage solution in Dracarys is far more integrated and a lot “closer to the metal” so to speak compared to the Radeon Pro SSG. Which is more like a graphics card and a PCIe SSD integrated into one expansion card. Where there are a few hoops that have to be jumped over for the Fiji GPU to access all of the available storage space on the SSD, rather than having it as a direct pool of memory. In that respect alone, the new Vega 10 powered Dracarys board is very different. Another key difference is the new 20+ FP16 teraflops VEGA 10 graphics engine which has more than double the graphics horsepower of Fiji In the Radeon Pro SSG. Suffice to say, Dracarys is dramatically more potent.

ADVERTISEMENT
This fiery new monster is set to launch in the second half of next year. Some units will be available at an earlier date, for a cool $10,000. Dracarys is exquisitely unique that it has effectively no competition, simply because there’s no other product like it. Which means that AMD can charge for it as much as its customers are willing to pay.

Although it’s important to point out that the final product will cost significantly less than the development kits. We would still imagine however, that the end-product won’t come cheap by any means. Also, depending on how well the board is received, we could see AMD pursue penetration across more market segments. Which would bring the price down on these kinds of extremely high capacity graphics products. Perhaps making them more commonplace, especially in the high-end video production market."

 

[zinfamous]

...Dracarys?

.....

 

[kawi6rr]

Is that 1TB+ of HBM?

 

[kraatus77]

http://wccftech.com/amd-dracrays-vega-10/

WCC says:

"We’ve just learned about AMD’s craziest graphics product to date. A ten thousand dollar behemoth powered by Vega 10 & 1TB+ of memory & we’re going to share all the details with you today, exclusively.

This next generation mini super-computing fire-breathing graphics monster is code named “Dracarys”. The name, incidentally, translates to “dragon fire” in High Valyrian, a language in Game Of Thrones lore. As of today, we’re told, that early Dracarys prototypes are up and running and ready for primetime. In fact, our sources tell us that we’ll get a sneak peak at this Vega 10 powered graphics board as early as next month at Supercomputing 2016.

Vega 10, The Pulsating Heart Of AMD’s New Fire Breathing Monster
AMD’s next generation Vega 10 GPU will power this unique board which will also be equipped with an excess of one terabyte of directly accessible on-board graphics memory. Represented in 16GB, interposer linked, 2.5D stacked, second generation high bandwidth memory in addition to an on-board solid state device. So, in many ways, Dracarys is quite similar to AMD’s recently announced Radeon Pro SSG. Which features a Fiji GPU as well as an integrated SSD storage option, expandable via an M.2 slot.

However, the storage solution in Dracarys is far more integrated and a lot “closer to the metal” so to speak compared to the Radeon Pro SSG. Which is more like a graphics card and a PCIe SSD integrated into one expansion card. Where there are a few hoops that have to be jumped over for the Fiji GPU to access all of the available storage space on the SSD, rather than having it as a direct pool of memory. In that respect alone, the new Vega 10 powered Dracarys board is very different. Another key difference is the new 20+ FP16 teraflops VEGA 10 graphics engine which has more than double the graphics horsepower of Fiji In the Radeon Pro SSG. Suffice to say, Dracarys is dramatically more potent.

ADVERTISEMENT
This fiery new monster is set to launch in the second half of next year. Some units will be available at an earlier date, for a cool $10,000. Dracarys is exquisitely unique that it has effectively no competition, simply because there’s no other product like it. Which means that AMD can charge for it as much as its customers are willing to pay.

Although it’s important to point out that the final product will cost significantly less than the development kits. We would still imagine however, that the end-product won’t come cheap by any means. Also, depending on how well the board is received, we could see AMD pursue penetration across more market segments. Which would bring the price down on these kinds of extremely high capacity graphics products. Perhaps making them more commonplace, especially in the high-end video production market."

Another BS from clickbait crap site.

 

[HurleyBird]

This one seems a lot more believable than most of their articles, to be honest. The majority of their "finds" are just common knowledge mixed with speculation and hype, or regurgitating theories from others. This story, on the other hand, at least has the benefit of being specific and novel, while also seeming very plausible.

 

[biostud]

Is that 1TB+ of HBM?

No 16Gb HBM + 1TB SSD

 

[Headfoot]

This one seems a lot more believable than most of their articles, to be honest. The majority of their "finds" are just common knowledge mixed with speculation and hype, or regurgitating theories from others. This story, on the other hand, at least has the benefit of being specific and novel, while also seeming very plausible.

I agree. I don't trust WCCFTech but this smells like something they picked up wholesale from someone more credible

 

[gamervivek]

I agree. I don't trust WCCFTech but this smells like something they picked up wholesale from someone more credible

AMD had been shipping Fiji with SSD to India at the start of this month.

PRINTED CIRCUIT BOARD ASSEMBLY (VIDEO/ GRAPHIC CARD)DRACARYS FIJI/SSD P/N:102-D02702-00 (FOC)

https://www.zauba.com/import-printed+circuit+board+assembly/hs-code-84733030/ip-INHYD4-hs-code.html

I thought it was a Vega card earlier.

 

[Head1985]

about clock speed debate at same process...1050Ti can reach 2Ghz(almost) at 14nm..So polaris low clock is not because 14nm.
Polaris 1350-1400Mhz max and pascal 2Ghz at same process.
http://videocardz.com/63940/amd-quietly-launches-radeon-rx-470d-for-chinese-market

 

[Bacon1]

about clock speed debate at same process...1050Ti can reach 2Ghz(almost) at 14nm..So polaris low clock is not because 14nm.
Polaris 1350-1400Mhz max and pascal 2Ghz at same process.
http://videocardz.com/63940/amd-quietly-launches-radeon-rx-470d-for-chinese-market

They are completely different architectures. Dunno why everyone thinks that AMD would be @ 2ghz when Hawaii/Fury was @ ~1000 vs Maxwell's ~1400.

Polaris went from 1000s -> 1300s (30%) and Pascal went from 1400s -> ~1800-1900s (~30-35%).