AMD Vega 10, Vega 11, Vega 12 and Vega 20 Rumors and Discussion

[iMacmatician]

I see it the same Pony, but this should be added: unlike Apple products, some faulty graphics chips can be salvaged in cut-down products.

The A9X in the 9.7" iPad Pro has a cut-down memory bus and the A5 in the Apple TV (3rd gen) has one CPU core disabled.

But in general, for a given chip Apple avoids large differences in specs.

 

[w0lfram]

It would cost too much. The Vega 20 is targeted squarely at the high end AI market because they're the ones that can afford to pony up for a giant chip built on a just released node. The reason the 20 will cost so much is the same reason Nvidia didn't delay a few months to put Turing on 7nm. When new nodes come out error in building the chips is high.

If the error crops up in a non critical area you can "bin" chips, disable that part (a shader core, etc.) and sell it at a lower price. But if an error crops up in a critical area it could kill the whole chip. If your building relatively tiny chips, such as Apple's A12 at 83mm (squared), and a critical error shows up on average of once every 400mm, then a little over 20% of A12 chips would be defective. Bad but not the end of the world for trillion dollar company Apple. Now a Vega 64 is 486mm (squared) in die size. Even squeezing that to 7nm, under half size, that's still every other chip being defective, that's a lot of wasted money.

And these are just example numbers, which don't take into account things like transistor cost not going down for nodes anymore. So even on 7nm, if Vega 20 is just double a Vega 64, then it costs double to make it. Meaning there's no way AMD is hitting a reasonable price point for gamers with it, not including lost chips. As a time goes on in a node each silicon foundry refines its process and bring errors down, so it's conceivable we'll see Navi as a much smaller, mainstream chip ($200-$300) by the middle of next year, but Vega 20 isn't being released at any reasonable price point this year.

7nm does not cost double, to make.
The smaller process yields much more chips per wafer... & the actual "cost" per usable gpu, is less. They can afford to sell it for less, but are geared for higher profits to pay off the premium for buying up (contracting) the 7nm nodes first. (7 months before nvidia's 7nm gaming gpu.)

7nm spin would allow for a GPU chip smaller than a Vega64 chip, with 112 CU's and GDDR6 memory, & most likely using less than 300w. Along with USB-C.


Understand, nobody is talking about Vega20, per se... we are talking about co-spin of that 7nm chip, but geared more toward gamers @ 7nm. AMD's uarch is already been proven to be quite modular.

What is stopping AMD from releasing a "gamer stepping" of the Vega20 by the holidays on 7nm..?



Specially when infinity fabric for gaming was mentioned by the end of the year?

 

[CarstenS]

What is stopping AMD from releasing a "gamer stepping" of the Vega20 by the holidays on 7nm..?

While I agree that AMDs µarch is quite modular, for this particular thing I guess capacity is the key. Why go the extra mile of designing and taping out an extra chip only to sell it for far less than the chip you already have in the pipeline targetted at the professional market? Even though Radeon MI and Pro are much cheaper than their Nvidia counterparts, they are still much more expensive than equivalent RX Radeon branded cards.

Additionally, AMDs marketing would have had to drop little hints for that upcoming products at the time of Turing's introduction at the latest, because (nearly) every gamer upgrading right now is diminishing the already small TAM by one. And they should have an easy time convincing people to wait until 2018's holiday season if they play the pricing angle clever.

 

[DmitryKo]

What is stopping AMD from releasing a "gamer stepping" of the Vega20 by the holidays on 7nm..?

Limited supply?

It just looks like AMD has always planned to commence mass production of Navi on the 7nm node, because it's the architecture for next-gen consoles.

 

[Kaotik]

It just looks like AMD has always planned to commence mass production of Navi on the 7nm node, because it's the architecture for next-gen consoles.

I doubt consoles have absolutely anything to do with what AMD releases or doesn't release for PC gamers

 

[DmitryKo]

The majority of RX Vega engineering team was reassigned to Navi, which was given priority over high-end desktop products in a bid for PS5, according to this report by Forbes - therefore Navi would end up as a mass-market, mid-range GDDR6 part with Infinity Fabric for Zen2 APU interconnect only, which is basically where 7nm Vega10 would stand (whereas Nvidia is actively researching high-end multi-chip GPU designs).

 

[Ike Turner]

The majority of RX Vega engineering team was reassigned to Navi, which was given priority over high-end desktop products in a bid for PS5, according to this report by Forbes - therefore Navi would end up as a mass-market, mid-range GDDR6 part with Infinity Fabric for Zen2 APU interconnect only, which is basically where 7nm Vega10 would stand (whereas Nvidia is actively researching high-end multi-chip GPU designs).

I wouldn't put too much faith into Forbes "contributors" or any other honky report coming from the likes of wccftech who "strangely" posted exactly the same story on the same day but quoting their own "sources".

 

[DmitryKo]

We shall find out on January 9, 2019: AMD President and CEO Dr. Lisa Su to Keynote at CES 2019

 

[w0lfram]

While I agree that AMDs µarch is quite modular, for this particular thing I guess capacity is the key. Why go the extra mile of designing and taping out an extra chip only to sell it for far less than the chip you already have in the pipeline targetted at the professional market? Even though Radeon MI and Pro are much cheaper than their Nvidia counterparts, they are still much more expensive than equivalent RX Radeon branded cards.

Additionally, AMDs marketing would have had to drop little hints for that upcoming products at the time of Turing's introduction at the latest, because (nearly) every gamer upgrading right now is diminishing the already small TAM by one. And they should have an easy time convincing people to wait until 2018's holiday season if they play the pricing angle clever.

Capacity..? (AMD's, or TSMC's..?)

Are you asking why AMD would use VEGA20's process (7nm) and tape out a gaming version of it..? Because it would be the fastest video card on the market, using less power and put AMD nearly 8 month ahead of Nvidia's 7nm GPUs.

Additionally, Dr Su and several other engineers have already mentioned it.



Why continue to make Vega56, or 64's, when u can make them on a new node for the same price with 30% more transistors..?

 

[CarstenS]

TSMC's. The rest follows naturally.

 

[entity279]

wait what? someone here's under the impresion vega 20 is 7 nm?Or that we can have any 7 nm gpus in 1-2 months?

don't really get what @w0lfram's on about

 

[Deleted member 13524]

Are you asking why AMD would use VEGA20's process (7nm) and tape out a gaming version of it..? Because it would be the fastest video card on the market, using less power and put AMD nearly 8 month ahead of Nvidia's 7nm GPUs.

Instinct and Pro cards sell for several thousands of dollars. Consumer Radeons sell for under $1000.

someone here's under the impresion vega 20 is 7 nm?

You're under the impression it's not?

https://www.anandtech.com/show/12910/amd-demos-7nm-vega-radeon-instinct-shipping-2018

Or that we can have any 7 nm gpus in 1-2 months?

This is Lisa Su holding a 7nm Vega 20 in her hand. From 4 months ago.

 

[Rootax]

For all we know, it's "papier-maché" *with George Costanza voice*

They said no Vega 20 7nm for gaming. I get that on paper, It could be a good upgrade over Vega 64, but I don't think they plan on pushing this uarch for gaming, and their focus for gaming is on launching Navi now.

 

[Kaotik]

wait what? someone here's under the impresion vega 20 is 7 nm?Or that we can have any 7 nm gpus in 1-2 months?

Yes, and it's been known for quite some time already as pointed out by @ToTTenTranz

 

[eastmen]

For all we know, it's "papier-maché" *with George Costanza voice*

They said no Vega 20 7nm for gaming. I get that on paper, It could be a good upgrade over Vega 64, but I don't think they plan on pushing this uarch for gaming, and their focus for gaming is on launching Navi now.

well I hope navi is early next year , they need something out soon

 

[Deleted member 13524]

well I hope navi is early next year , they need something out soon

With a Polaris 30 / RX590 apparently coming out in November, I doubt Navi is coming before late Q2 2019.

I think it's more likely for them to launch a consumer Vega 20 with 2 or 3 HBM2 stacks and lower FP64 ratio just to compete at the high-end for half a year, than to release Navi earlier than H2 2019.

 

[Alexko]

For all we know, it's "papier-maché" *with George Costanza voice*

They said no Vega 20 7nm for gaming. I get that on paper, It could be a good upgrade over Vega 64, but I don't think they plan on pushing this uarch for gaming, and their focus for gaming is on launching Navi now.

Did they? All I remember was a statement that there would be 7nm GPUs for gaming—well, duh—but without confirming or denying whether Vega 20 was one of those 7nm gaming GPUs. This sort of hints at Navi being the only 7nm gaming architecture (along with its replacement, of course), but it's not quite the same thing as stating outright that Vega 20 would be restricted to the professional market.

Besides, we could see another semi-pro Frontier Edition card, this time based on Vega 20, that could perform quite well in games, albeit at a prohibitive price. Then again, the Vega FE was released at a lower price than the 2080 Ti, so where's the line between a gaming card and a (semi-)pro one these days?

If you want my opinion, back when AMD first made this statement, they probably didn't know whether Vega 20 would be offered as a gaming SKU, because the competitive situation at the time of its release wasn't known; there were probably a bunch of unknowns about the final product and its production (performance, yields, capacity, etc.), and I'd wager that some of these things still aren't known very precisely. So my guess is that AMD will release Vega 20 to the pro market, and then evaluate whether it also makes sense as a gaming card—it could go either way.

 

[eastmen]

With a Polaris 30 / RX590 apparently coming out in November, I doubt Navi is coming before late Q2 2019.

I think it's more likely for them to launch a consumer Vega 20 with 2 or 3 HBM2 stacks and lower FP64 ratio just to compete at the high-end for half a year, than to release Navi earlier than H2 2019.

That's a shame , I like my vega 56 but would not mind more power. But with nvidia disappointing and amd not showing up I will have to wait till next fall I guess.

 

[Rootax]

well I hope navi is early next year , they need something out soon

Don't forget they skip the previous high end war and stayed with Fiji, and only competing in the mid range with Polaris. They're not on the same timing than nVidia...

 

[Deleted member 13524]

That's a shame , I like my vega 56 but would not mind more power. But with nvidia disappointing and amd not showing up I will have to wait till next fall I guess.

A consumer Vega 20 could be pretty competitive if it clocks north of 2GHz and has 12GB of 2.4 Gbps HBM2.