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

So, Engadget put up video about AMD tech at CES 2018, and at 1m44s he talks about AMD announcing their first 7nm product, a Radeon Instinct card based on Vega. Seems like, pretty big news if true? Maybe this is Vega 20?

If they can already announce real 7nm products, especially such a big chip, this seems to bode well for 7nm progress.

Can't post links, but here is the video ID: /watch?v=Z7_zunLzrk8&t=1m40s
 
So, Engadget put up video about AMD tech at CES 2018, and at 1m44s he talks about AMD announcing their first 7nm product, a Radeon Instinct card based on Vega. Seems like, pretty big news if true? Maybe this is Vega 20?

If they can already announce real 7nm products, especially such a big chip, this seems to bode well for 7nm progress.

Can't post links, but here is the video ID: /watch?v=Z7_zunLzrk8&t=1m40s

This is 100% TSMC 7nm.
 
So, Engadget put up video about AMD tech at CES 2018, and at 1m44s he talks about AMD announcing their first 7nm product, a Radeon Instinct card based on Vega. Seems like, pretty big news if true? Maybe this is Vega 20?

If they can already announce real 7nm products, especially such a big chip, this seems to bode well for 7nm progress.

Can't post links, but here is the video ID: /watch?v=Z7_zunLzrk8&t=1m40s
I was able to play the video, then I went back and it was gone! I guess AMD is not you happy about the early coverage...

Sent from my SM-G950U using Tapatalk
 
Wishful thinking. There are a lot of people who hate GloFo, and think TSMC would make everything better. But if this thing would be released in the next two month, or so, it may actually be TSMC 7nm - I don't think GloFo ramped up their 7nm production yet.
 
Wishful thinking. There are a lot of people who hate GloFo, and think TSMC would make everything better. But if this thing would be released in the next two month, or so, it may actually be TSMC 7nm - I don't think GloFo ramped up their 7nm production yet.
No, they haven't ramped it up yet, but the way it was worded I didn't get the impression it would be available right after CES either which gives them some time (though I suppose it would be a bit too early to announce product that wouldn't be available 'till H2 which is when GloFo is supposed to have 7nm in mass production)
 
GloFo 7nm should be too late for any 2018 product launch. At best manufacturing start.

TSMC 7nm is the only candidate here unless GloFo ends up accelerating their roadmap or TSMC falls behind schedule.
 
GloFo 7nm should be too late for any 2018 product launch. At best manufacturing start.

TSMC 7nm is the only candidate here unless GloFo ends up accelerating their roadmap or TSMC falls behind schedule.
Products have been released before around the same time (give or take month or two) as mass production starts. If GloFos 7nm is early 2H instead of late 2H it's plenty early for 2018 launches
 
No, they haven't ramped it up yet, but the way it was worded I didn't get the impression it would be available right after CES either which gives them some time (though I suppose it would be a bit too early to announce product that wouldn't be available 'till H2 which is when GloFo is supposed to have 7nm in mass production)

Mi25 was announced in december 16 and available sept. 17. I'll expect the same with this 7nm chip. End Q3/Early Q4 for availability. So it might even be possible for gf.
 
Mi25 was announced in december 16 and available sept. 17. I'll expect the same with this 7nm chip. End Q3/Early Q4 for availability. So it might even be possible for gf.
if its just vega with more shaders it shouldn't need the same amount of time in the oven so to speak. I would actually expect late Q2 early Q3 for this chip and hopefully Navi in early 2019. My vega has been rocking it so i will most likely skip vega 20 unless its a big enough increase and mining is still profitable. My vega has already paid for itself
 
if its just vega with more shaders it shouldn't need the same amount of time in the oven so to speak. I would actually expect late Q2 early Q3 for this chip and hopefully Navi in early 2019. My vega has been rocking it so i will most likely skip vega 20 unless its a big enough increase and mining is still profitable. My vega has already paid for itself

It's a 7nm chip. Of course it will take a lot of time and if the specifications rumoured are true, then this is amds take on a nearly pure compute chip. Anyway you won't get this stuff in the consumer space so fast. It's ok to have a expensive low yielding chip at process start in the pro segment, but the production costs will probably be too high for consumer.
 
That’s exactly what I was trying to say. IMHO, everything that COULD come from the drivers in existing software WOULD have already arrived. My position that IF we were to see big gains going forward, it will have to be via work on developer’s part, whether it be in form of patches to existing software or resource investment into projects still in development. IF it’s possible at all.


I do not think you understand how 4k gaming works. Or why AMD is holding back.

Secondly, I am not sure if you are being facetious but AMD's track record on RX Vega drivers since release, have pretty decent gains for just 6 months? What we are waiting to see, is a flurry of true DX12 games that use standard API's, that all game studios can take advantage of.

The technicalities of RX Vega's feature set are well known by now. When supported, Vega will get a boost.


Oddly, what many don't want you to focus on, is that fact that Volta in the future will also be utilizing DX12 fully too, in Games. It is hypocritical to be skeptical about RX Vega feature set and Game Developers using them, if you are not worried about Volta's.

Additional performance gains will come with the help from developers. It is a collaboration.
 
I do not think you understand how 4k gaming works. Or why AMD is holding back.

Please, DO enlighten the unwashed.

Secondly, I am not sure if you are being facetious but AMD's track record on RX Vega drivers since release, have pretty decent gains for just 6 months? What we are waiting to see, is a flurry of true DX12 games that use standard API's, that all game studios can take advantage of.

The technicalities of RX Vega's feature set are well known by now. When supported, Vega will get a boost.

So your adamant disagreement with my statement that future performance gains will be driven by explicit developer support consists of poorly worded declaration that future “boost” will come with ... developer support for Vega features?


Oddly, what many don't want you to focus on, is that fact that Volta in the future will also be utilizing DX12 fully too, in Games. It is hypocritical to be skeptical about RX Vega feature set and Game Developers using them, if you are not worried about Volta's.

Additional performance gains will come with the help from developers. It is a collaboration.

I don’t think you understood what I wrote, since my position is fairly clear: “Magic Drivers” that are going to increase the performance by 20-to-infinity percent in existing software are not happening; furtive increases, if such, will come with explicit developer support. You seem to be disagreeing with me while saying stating essentially the same thing.
 
Last edited:

Well the Hades Canyon NUC is a consumer SFF desktop with a Vega GPU.

Regardless, we already knew that a smaller Vega for desktop wouldn't make much sense, given that Polaris still gets the job done on that price range and uses cheaper GDDR5.
Maybe AMD could phase-out Polaris and make a smaller Vega with GDDR6, but that seems to be what Navi will be about come 2019.


I'm more curious to know if there will be an actual discrete version of Vega M for laptops or if Intel is taking exclusivity on those chips.
Anandtech seems to heavily suggest there are discrete Vega M products coming up in 2018, other than the Vega M inside Kaby Lake G:

Anandtech said:
On the graphics side, AMD’s roadmap was a little more conservative. With the launch of Vega 56 and Vega 64 in the middle of 2017, as well as the Radeon Instinct line, AMD’s focus will be on Radeon RX Vega Mobile discrete GPUs.



These mobile-class discrete GPUs are optimized for z-height, coming in at 1.7mm, but will also use HBM2 (high bandwidth memory) on an interposer, indicating an update on AMD’s prowess on interposer technology that previously has been somewhat bulky. The purpose here for Radeon RX Vega Mobile is to move discrete graphics into mobile and small form factor systems that are currently occupied by mid-range NVIDIA graphics. HBM2 is still a premium product, so we expect that when they come to market in Q2, they will aim for thin and light systems that also want to support VR-like technologies.

HBM+Interposer means a cost that is only justified for premium laptops.

I'm guessing Kaby Lake G takes the mantle for 65-100W laptops, and then for higher-end (or just larger) laptops, Vega M can be paired up with the 6-core Coffee Lake H to compete with the GTX 1060.

And if that's true, I wonder if this new Vega M chip with 1 HBM stack isn't actually just the same as we''re seeing in Kaby Lake G, but with higher core clocks, more CUs activated (say 32 CUs) and higher HBM clocks.
 
Seems like VEGA12 == "Ultrathin Vega Mobile" == Kaby G GPU

3s2Pm4F.jpg


Via Videocardz.
 
Back
Top