AMD mentions Vega Enthusiast GPUs to be released H1 2017
An intersting spot, in a recent AMD investor presentation of August 2016 there is a mention of 'Vega' for the enthusiast market in the slides, with a launch date set at 1H 2017. That could be Vega 11.
1H 2017 would mean first half of 2017. Now we know there will be multiple Vega (Vega 10 and Vega 11) based products so that is an interesting spot alright as AMD originally was supposed to launch their high-end class VEGA (codename) product in early 2017, as it now seems that launch is being moved forward towards October.
"Vega10" is lined up to replace the 390/490 series. "Vega11" will replace the FIJI / Fury parts. Rumors right now indicate that Vega10 has 4096 stream processors and could compete with the GeForce GTX 1080 and GTX 1070. Vega11 (thus that launches 2017 1H would feature 6144 stream processors and would be lined against the GP100 big pascal GPU (think Titan).
The new info on VEGA 11 can be spotted in this presentation.
AMD decided to pull the launch of its next big silicon, Vega10 from early-2017 launch to October 2016 (that is still a rumor though). Vega10 would be the GPU that replaces the Hawaii/Grenada series, perhaps with GDDR5X memory ? If so, Vega11 could be the enthusiast class HBM2 memory product that releases in 2017.
Senior Member
Posts: 550
Joined: 2016-05-24
So no competition and price fall for G1070 and G1080 for new 8 months, or until Nvidia will deliver even more expensive HBM models..
Senior Member
Posts: 2141
Joined: 2011-01-05
Great... prices will remain the same for the next year. I guess the wait will continue until 4K HDR screens become mainstream before we some of us upgrade.
Senior Member
Posts: 3174
Joined: 2006-04-25
So, AMD has no high end GPU option anytime soon. That is not good for prices nor for pushing to make even better tech.
Senior Member
Posts: 11158
Joined: 2004-05-10
Big different there. And since when has AMD ever shown, like ever, that they are even close to Nvidia when it comes to watt/perf?
5850/5870 vs Fermi.
Posts: 33
Joined: 2016-08-25
Those features aren't selling points for everyone. Personally, I really couldn't care less about PhysX. I haven't touched Inspector since I move away from SLI and Gameworks support isn't a game breaker either. Every cross platform game is going to have to be optimized for AMD hardware anyway since XBox One and PS4 both use AMD hardware. If the devs remove any optimizations specific to AMD hardware from cross-platform games, AMD has legal recourse against them unless the devs can prove that such optimizations will break the game. How many devs are going to risk it?
I agree those features matter little, what matters is performance and cost mainly. Power consumption matters only insofar as temperatures do for most people.
@ $1500 +13% tax CND I don't care if the TITAN X had come out in 2015
http://www.ncix.com/detail/pny-geforce-gtx-titan-x-76-128292.htm
no sale here so buy as many as you can before they go oos where you live.
and for me money is not a issue ,
By summer of next year the next line of GPUs will be rolling out with the X70 part probably matching a stock Titan X. So by the time Vega 11 becomes relevant, it will be a lower high-end part like the 1070.
I dont know why but i got the feeling that the small vega will be released in October along with Battlefield 1 Launch "its just a hunch" , but even if what you said reveals to be true , it aint that bad "an overclocked titan x" one year later
with a lil difference , i doubt that big vega will cost 1200$.
maybe within the first half they will release the entire lineup.
Yeah it would be nice to see Vega 10 in October, should tell us how Vega 11 will perform as well
Also worth noting that while AMD can always bring the fight in terms of pricing, they're in no position to do this.
With Polaris on 14nm GloFo many of us thought AMD would finally be able to sell smaller (less expensive) dies at the same price as NV's larger offerings.
14nm GloFo vs 16nm FF should be around a 10% size advantage.
GP106 is smaller than Polaris 10; 200mm^2 vs 232mm^2 on P10.
So 180mm^2 vs 232mm^2 in GloFo terms.
This means, barring large architectural changes, that Vega will not only be more expensive because it's bigger, but also because of HBM.
So while it may make people happy to see them undercutting NV's offerings, it's not sustainable imo.
That said if they are able to clock 20% higher without outrageous power consumption increases it should be fine.
Vega 10 is already confirmed to be essentially Fiji; 64 CUs for 4096 shaders.
50% die size increase seems likely so we are talking around 360mm^2