• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

[Rumor, Tweaktown] AMD to launch next-gen Navi graphics cards at E3

Page 71 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
Status
Not open for further replies.
Shivansps said:
Navi 10 IS NOT Vega 56/64 replacement, it is Polaris 10 replacement. Period.
Click to expand...
This is a game of semantics.

For people building new rigs, its pretty much an exact replacement. about 400 hundred bucks and better performance. So if I am shopping for a GPU and I am looking for that price level I now would look at Navi while before I might look at Vega.

For AMD it is a replacement in product stack for Polaris. There's no definition of "replacement" so there's no "period" after a statement like that.
 
swilli89 said:
This is a game of semantics.

For people building new rigs, its pretty much an exact replacement. about 400 hundred bucks and better performance. So if I am shopping for a GPU and I am looking for that price level I now would look at Navi while before I might look at Vega.

For AMD it is a replacement in product stack for Polaris. There's no definition of "replacement" so there's no "period" after a statement like that.
Click to expand...

if Navi 10 is a replacement product stack for Polaris as you said, IT IS Polaris replacement, witch they are overpricing it to Vega levels, there is no room for "buts" here. Semantics is what some people is using to try to justify this.

AMD thinks they can get away with this, if they do, both AMD and Nvidia wins, and we losse, choose your side. Hang on, i belive you said you own AMD stock, well there is nothing more to say if thats the case.
 
Shivansps said:
Navi 10 IS NOT Vega 56/64 replacement, it is Polaris 10 replacement. Period.
Click to expand...
Because you said so? 😉

AMD is placing it as Vega replacement. In every single promotional material it is placed agains Vega 56/64. And yet, you say it is placed as Polaris replacement. What do you have to back that up?
 
Shivansps said:
if Navi 10 is a replacement product stack for Polaris as you said, IT IS Polaris replacement, witch they are overpricing it to Vega levels, there is no room for "buts" here. Semantics is what some people is using to try to justify this.

AMD thinks they can get away with this, if they do, both AMD and Nvidia wins, and we losse, choose your side. Hang on, i belive you said you own AMD stock, well there is nothing more to say if thats the case.
Click to expand...
"get away with"?

Look I dislike the pricing as well, but AMD's balance sheet needs help and they need revenue and margins. They are undercutting NVIDIA and that's all they have to do right now. NVIDIA is the one with an old product. Let them discount and force AMD to reduce pricing.

https://twitter.com/x/status/1139607746397888512

Anandtech's own RT said it doesn't even have classification.

Big Navi will probably be 80 CU and by the time it drops will land at $699.
 
Last edited:
If Vega 56/64 disappear from the consumer market then we can reasonably conclude that Navi is a Vega 56/64 replacement.
 
swilli89 said:
"get away with"?
Look I dislike the pricing as well, but AMD's balance sheet needs help and they need revenue and margins.
Click to expand...
No they dont need any help.They just milking us just like nvidia..Its freaking corporation.Stop defending price increase.They simple doubled price vs polaris thast all.If you keep buying navi at 450usd you can be sure next midrange will cost 600usd.
 
Last edited:
guachi said:
If Vega 56/64 disappear from the consumer market then we can reasonably conclude that Navi is a Vega 56/64 replacement.
Click to expand...
I promise you AMD is no longer ordering Vega wafers from GlobalFoundries. Each chip is 500mm2 and requires HBM memory and exotic substrate packaging. Why would AMD choose to manufacture this when a 251mm2 Navi with normal GDDR6 memory interface is able to be sold at the same or higher price and is much more attractive to consumers?

Vega is past EOL once Navi is out. Actually for that matter 7nm Vega, Radeon VII is done as well. For now its just Polaris and Navi
 
Last edited:
RTX 2070 is 970 successor but was priced at 600 instead of 330 like 970,similarly 5700xt is a rx580 successor priced at 450 instead of 240. More expensive 7nm process,a little higher margins should still keep it under 300 but here we are prices increased to almost double in very short time. You cannot call this a Vega 64 replacement.You call this price gouging.
 
mohit9206 said:
RTX 2070 is 970 successor but was priced at 600 instead of 330 like 970,similarly 5700xt is a rx580 successor priced at 450 instead of 240. More expensive 7nm process,a little higher margins should still keep it under 300 but here we are prices increased to almost double in very short time. You cannot call this a Vega 64 replacement.You call this price gouging.
Click to expand...

Nah. Nvidia is price gouging like they have done for years. AMD is offering more performance for less money, as usual. You need to be crusading against nvidia to lower their prices. Especially since their cards have been on the market for 6 months, prices should have already dropped.
 
piesquared said:
Nah. Nvidia is price gouging like they have done for years. AMD is offering more performance for less money, as usual. You need to be crusading against nvidia to lower their prices. Especially since their cards have been on the market for 6 months, prices should have already dropped.
Click to expand...
People are not voting with their wallets so it seems this will continue. Add a useless feature like RTX and charge double and people still buy. People paid 600 for 2070 its kind of unbelievable.
 
mohit9206 said:
Despite that,leaked price of RTX 2080 Super is $799.
Click to expand...

Hard to judge without performance numbers. maybe it's based on a cut TU102. Rumors also say 2070 super will be on TU104 and not 106 anymore which could make it more viable for RT.

As someone working in data science the potential benefits for Deep learning with the tensor cores does have some value.
 
Head1985 said:
Looks like everyone and their mother just want milk PC gamers lately.NV/AMD price fixing and overpriced x570 boards on top of that.I am surprised that AMD didnt doubled prices also with zen 3000.
Click to expand...

PC gamers? No, its all of them. Not just gamers, not just PC, not just tech.

mohit9206 said:
Add a useless feature like RTX and charge double and people still buy.
Click to expand...

The current Turing RTX cards are surely overpriced, but Ray Tracing is not useless. Next year some BIG AAA titles will be coming with Ray Tracing support. And since Ray Tracing cores are new, I expect gains with RT enabled to be greater than with conventional games with the next gen cards. 25% faster in regular games, 40-50% faster with RT enabled?

If Ray Tracing offers benefits to developers like better quality at lower development effort, in the long run it'll be better for all of us.

The questionable feature is called DLSS.
 
swilli89 said:
I promise you AMD is no longer ordering Vega wafers from GlobalFoundries. Each chip is 500mm2 and requires HBM memory and exotic substrate packaging. Why would AMD choose to manufacture this when a 251mm2 Navi with normal GDDR6 memory interface is able to be sold at the same or higher price and is much more attractive to consumers?

Vega is past EOL once Navi is out. Actually for that matter 7nm Vega, Radeon VII is done as well. For now its just Polaris and Navi
Click to expand...
AMD is not ordering Vega for DIY consumer market. But they are still ordering those wafers for OEM needs.
 
I know it is not the thread to talk about it but, leaked prices of Super RTX GPUs are pretty much... disappointing.

But kinda show in what world we live in.
 
swilli89 said:
I promise you AMD is no longer ordering Vega wafers from GlobalFoundries. Each chip is 500mm2 and requires HBM memory and exotic substrate packaging. Why would AMD choose to manufacture this when a 251mm2 Navi with normal GDDR6 memory interface is able to be sold at the same or higher price and is much more attractive to consumers?

Vega is past EOL once Navi is out. Actually for that matter 7nm Vega, Radeon VII is done as well. For now its just Polaris and Navi
Click to expand...
AMD themselves alluded to a split between gaming and compute designs going forward. Vega by this definition, cannot be dead, unless you think they're abandoning the heavy workload compute market. As I see it, once they have both product lines, they might as well cross sell a few.

Who here still thinks that Radeon VII was not an opportunistic product for gamers.
 
NostaSeronx said:
Navi is a Polaris successor. Vega's successor is getting phat boi Navi+ with 128 CUs and 2x HBM3(N6) or 4x HBM3(N3).
Click to expand...
Even at 7nm+, even using 4x32CU chiplets, 128 CU's would be insane, no 2 ways about it. Thats more 7nm silicon than Rome has total, with much larger logic dies (lower yields).

Just a short bit of logic die shrink math:

0.8x250mm2 = 200mm2 (7nm+ Navi 10).
0.8x200mm2 = 160mm2 (7nm+ Navi 10 with only 32 CU's).

Even if you strip away the GDDR memory controllers and duplicated media + display engines, they would probably still be twice the area of Zen 2 chiplets.
its basically around 560-640mm2 just on logic alone, leaving out the obligatory IO+C&C die (C&C = Command and Control, necessary to keep the package a singular entity to the OS).

I've probably fudged the math here somewhere, but that is still a rediculous amount of silicon on a relatively new process for high end chips - it would cost closer to 64 core Threadripper than any current AMD GFX card.
 
Last edited:
NostaSeronx said:
or 4x HBM3(N3)
Click to expand...
Also, N3 is 2021 at the earliest, it uses a brand new GAA device type called a Nano Sheet, which departs from the FinFET structure of the past few years, so it won't be smooth sailing, its not even as area efficient as FinFET at ISO pitch size, just more power efficient.

Supposedly there is a revised Nano Sheet device called a Fork Sheet, which perhaps could be used for 2nm, or maybe even 1.5 nm (IMEC called it a mid life improvement for Nano Sheet).
This Fork Sheet device should provide even better power and FinFET/FinFET+ density.
 
Last edited:
Status
Not open for further replies.

TRENDING THREADS

Back
Top