AMD: Speculation, Rumors, and Discussion (Archive)

[yuri]

It's not impossible it's a Vega-variant, given the current schedule

Who knows. Anyway, Vega still aims for 1H 2017.
Lisa Su on 30th Nov:

But we also have our Vega architecture and graphics that’s coming out for high performance graphics in the first half of '17.

 

[jacozz]

Probably just the launch of the new High-End AM4-motherboards.
Maybe a sneak peak of Vega? Remember, they did that a year ago with Polaris 11.
If it is a dual Polaris card, I hope they found a new way around the crossfire problem in a way the OS only see one gpu. somehow... MCM? nah... probably not.

 

[CarstenS]

It could just be X370 AM4 motherboards, bringing their chipset feature-set up to par with Intel.

With USB 3.1 gen2 already being deployed on AM4, AMD is even ahead of Intel this time.

 

[CSI PC]

Damn it why make it 1080p if everyone else is doing 1440p?
Took me like 10 minutes to find a 1080p result for a RX480...

In other tests, the "total average" seems to be very close to the "medium" results.
This doesn't seem to be a dual-P10 card. At least for the fact that the game is only detecting one GPU.

It is rather frustrating to find any separate benchmarks to cross reference the results, even other sites do not necessarily have the same settings or even mention what those settings are
That brings us on to comparing the results to the AoTS result for the unknown card.
I took an alternative approach and tried to identify the settings compared to the ones shown for the unknown card, and unfortunately it seems to me we may not be able to use any benchmark review as a comparison for the following reason.
Looking at another review site they show their internal preset 'High' setting has 2xMSAA, texture quality High, everything else matches that unknown card settings.
That unknown card is using 1xMSAA, not sure about texture quality.
Cheers

 

[ieldra]

Also worth noting the game version used in the 'leaked' benchmark is 1.45, whereas the current game version is 2.1.x I believe

 

[CSI PC]

Probably the closest reference we have is WhyCry running same setting albeit with a 6800k at 3.4GHz rather than a 5930k.

He used same Game version and same settings, his card is the MSI GTX1080 Gaming X.
WhyCry: http://videocardz.com/64475/are-those-radeon-rx-490-vega10-benchmarks-results

And the highest leaked result so far (found by gamervivek):

May try to find WhyCry's result in the benchmark table.

Cheers

 

[gamervivek]

Hopefully these are driver improvements and there's more to come.

 

[homerdog]

I seriously hope Vega is good so it forces NVIDIA to put out a GTX1075 at ~$300. The 1070 is a bit too crippled and expensive IMO.

Of course if Vega is that good I would consider it a purchase option. My last AMD card was the 7950 and that turned out to be a great buy at $200.

 

[CSI PC]

Hopefully these are driver improvements and there's more to come.

The score keeps increasing every day it seems,
so is that person overclocking the GPU and if so that does make it tricky to know the realistic range.
I really doubt the person has updated engineering drivers since the last result, these are usually pretty well controlled and do not happen as frequent as these benchmark updates in terms of what could get out of the controlled lab-network.
Cheers