AMD: Navi Speculation, Rumours and Discussion [2019-2020]

Status
Not open for further replies.
There will most likely be a node change, even when they didn't say specificly "N7+ EUV" they did say "enhanced 7nm" which can only mean either N7+ EUV or N7P, but since Navi 10 is already N7P it's unlikely they'd call that same node "enhanced 7nm"
Going from 7P to 7+ should net them 20% power gain at best so it's still another 30% to cover with architectural changes. Quite a lot considering that it was way less than that between GCN@7 and RDNA@7P.
 
Q4 seems to be too late for AMD to cause any damage to nV. Ampere will likely launch earlier and will achieve much better perf/W.
Seems likely, but remains to be seen. Hopefully, AMD will be able to do more than just compete with 2080 Ti (a then >2 year old product) in the enthusiast space.
 
Very well.
They didn't seed those Zen mgmt people into RTG just because.
Navi3x series will also kick ass, since it has specific partner-derived goal.
Well, let's hope that this will be the case. It's long time overdue for AMD GPUs to fight NV on perf/watt again.
 
TSMC has three high-level versions of its 7nm process:
  • N7, which is the basic initial version using ‘DUV’ only tools (so no EUV),
  • N7P, which is the second generation version of N7 which is also only DUV
  • N7+, which is an EUV version of N7 for a number of layers in the metal stack

In order to avoid confusion, AMD is dropping the ‘+’ from its roadmaps. In speaking with AMD, the company confirmed that its next generations of 7nm products are likely to use process enhancements and the best high-performance libraries for the target market, however it is not explicitly stating whether this would be N7P or N7+, just that it will be ‘better’ than the base N7 used in its first 7nm line.

at this time they wanted to clarify that AMD is not making any specific clarifications of which version of 7nm from TSMC it plans to use

https://www.anandtech.com/show/1558...7nm-7nm-for-future-products-euv-not-specified
 
Well that's a funny assed mixup, probably should've been expected at some point since "+" is just such a vague, overused marketing signifier.

Guess this means Zen 3 and RDNA2 could be on N7P or even 6nm for the former, depending on when it comes out. Makes financial sense too, taping out EUV stuff as such is supposed to be really, really expensive, and it's not like 7nm+ actually gave almost any improvement (hence why almost no one used it).

Edit- They also pretty much confirmed Arcturus is their separate, Vega derived compute GPU. It's all compute concentrated and even, apparently, cuts graphics parts for it. They're officially splitting it into "CDNA", and since it doesn't computer graphics specifically one wonders if "GPU" is totally a good acronym anymore. Source: https://www.anandtech.com/show/1559...a-dedicated-gpu-architecture-for-data-centers

Anyway, I suppose one should split it off into its own thread since it's official and not RDNA.
 
Last edited:
I thought vega20 and navi10 were on the same node.
If Vega 20 is N7P then they are, but not sure if it's ever been confirmed which Vega 20 is

edit:
Not sure if it was mentioned already, but Navi2X support DXR 1.1, AFAIK there's no confirmation yet whether Turing does that too or is it limited to 1.0
 
Hu, I was expecting some benchs/numbers. They just said "yeah it's coming".
I get it was for investors, but still...
 
Status
Not open for further replies.
Back
Top