AMD Polaris 10 GPU To Offer Near 980 Ti Performance For 299 USD?

If true nvidia new cards will need to be reduced by price, also Polaris will help gain good amount of market.

 

Now this is very interesting.

If this is true I can see AMD selling these GPU's like hot cakes and with their added support for drivers now too. Things are looking up for them.

Add in the fact that we should see the back of 4GB VRAM and we should see GPU's with no less than 6GB VRAM from now on unless its entry level.

This is how I see it going

R9 470/X = 4GB RAM

R9 480/X = 6GB RAM

R9 490/X = 8GB RAM

R9 FURY2 = 12GB RAM (HBM1 or 2 or GDDR5X).

 

The red capes are coming... The red capes are coming!

 

That's never ending process. By the time they release it new games will make them midbudget not capable of providing modern image quality with decent fps.

 

**** i was hoping for at least a 10~15% performance increase. My 980TI needs replacing... Its not enough at 3440x1440 21:9 :bang:

Truly hoping that the GP104-400 is a beefed up version of the GP104-200 '1080' c1:

Haha yeah said the samething about the org. Titan (bought one). Its a great card but now its not even midrange. At 1080P High it struggle's with Rottr...

 

Will this not mean having to reduce the price across their own range as well?

Would be great if it was true, but more often than not new cards end up being a lot more expensive than the leaks suggest.

 

That would be a super upgrade from my 290X

 

so?
the 1080/1070 will end up in similar performance and price range, do a little ocing and i get the same...

 

This is quite hard to believe. While I expected AMD to clock higher on 14nm. Even 1500MHz would not give us such performance in this price range.
14/16nm FF wafer costs around 2.5x as much as 28nm wafer. Due to increased density it makes cost per transistor on 14nm only around 1.2~1.5times higher (based chosen density).

But even with only 1.2times higher price/transistor and 50% increased clock. It would give GCN only 25% increased performance/$.
Is remaining 75% coming from Drastic GCN changes (lower transistor count required per ROP/TMU/CU or lower clock count required to do certain work)? Or GCN on 14nm clocks that damn high?
I do not think even both of those would grant such increase in performance.

But with 10nm, I would expect this. As 14/16nm are about much better energy efficiency than 28nm. But 10nm main advantage over 14/16nm is actual manufacturing cost.
(But, who knows how well will 10nm clock...)

 

Then it comes down to pricing again, which was a main weapon of AMD in the recent past.

 

Why would you need to OC it? If 1080 can't match 980Ti, it will be a pointless product from Nvidia and wasted development costs.

 

Wow. So amazing if true!
So... when is the release date for these Polaris? I did a quick Google search and most said shipping in April but..... 4 days till end of April!!
I am on the verge of buying a 390X/Fury/Fury X but truly does seem pointless if Polaris are literally around the corner. I am too impatient to wait for NVIDIAS Pascal in June/July! (?)

 

I certainly hope that's not the 490X. I'd expect that card to perform well over a 980ti.

 

Imo no, because even if true, the demand will be overwhelming huge, and shortage always leads to higher prices. So better preorder now!

 

Will they finally be OC:able?

 

Folks is a R9 480x 8GB GDDR5/X 256 Bits

R9 490/X is faster and will be at 384/512 bits 12GB/16GB GDDR5X $500/600

 

If it's half true, i sell my -Ti and grab two of those haha!

 

If true i'll be getting one for sure. Come on AMD, this year is looking better and better.

 

At these prices, it is time to live off instant ramen and water for a couple of months and get CFX of these! It is a good time to plan an upgrade.