AMD's own testing? Fury X vs GTX 980Ti

[Khvarrioiren]

According to some benchmarks, the 300 series is only ~10% faster than the 200 series.

 

We're talking about a manufacturing process decrease to 14nm and a new revision of GCN. It's very likely to gain 50% performance over the latest GCN revision.

[d1screet]

We're talking about a manufacturing process decrease to 14nm and a new revision of GCN. It's very likely to gain 50% performance over the latest GCN revision.

Where's the 50% gone to then?

[Khvarrioiren]

Where's the 50% gone to then?

 

That would be going from the HD 6000 series to 7000 series.

[Sam Z Man]

Even if this is true G-Sync monitors still trump FreeSync so I believe it's still in 980's favor.

Evidence? I'm interested to see where this is coming from.

[GrimNeo]

Well AMD was off by 10fps on that Far Cry 4 4k benchmark. It only got 44fps.

[Nater]

Evidence? I'm interested to see where this is coming from.

 

 

I somewhat take my statement back considering there is a 1440 144hz IPS Freesync monitor for $599.

[Sam Z Man]

I somewhat take my statement back considering there is a 1440 144hz IPS Freesync monitor for $599.

So your retracting your statement completely.

[Sammael]

I find it very weird to set AF to 0, since AF doesn't affect performance more than 1-2 FPS even at 4K.

 

 

It would be interesting to see someone test the cards with AF set to 4/8/16 at 4k on the furys, but turn off msaa in favor of smaa/fxaa to see what kind of performance delta would be had then.  If msaa is the bigger area that taxes amd gpus more heavily, then shifting the workload to their much larger shader array to do the aa seems like a decent trade off with closer to / better than 980ti performance.  Someone should test that.

 

I don't see using settings that achieve a similar effect on your card that better takes advantage of the particular mix of hardware as a bad thing necessarily, so long as each gpu maker / user gets to choose the proper settings and effects that will make their own hardware shine.

[sgloux3470]

 

Not to mention, the additional added desktop real estate you get with 1440p and 4K is excellent for productivity oriented tasks.

 

Less than you'd hope for, since you have to scale everything 150-175% anyway at 4K to be able to use it, unless you're okay with squinting.

 

Nothing beats 21:9 for productivity atm.

 

 

Well AMD was off by 10fps on that Far Cry 4 4k benchmark. It only got 44fps.

 

Different settings probably.

[marldorthegreat]

Actually we do. 14nm FinFet is already up and running, and just like ZEN will use that node, so will Greenland. 20 nm at TSMC was just too leaky and poor performing, but FinFet's takes care of that. Pascal will either be 16nm FinFet+ (at TSMC) or 14nm FinFet along with Greenland on Global Foundries, though NVidia has yet to land that contract.

Like 20nm was "Up and running" for this generation? And we were expecting ths line of cards last year.

[marldorthegreat]

 

Less than you'd hope for, since you have to scale everything 150-175% anyway at 4K to be able to use it, unless you're okay with squinting.

 

Nothing beats 21:9 for productivity atm.

 

 

 

Different settings probably.

 

They didnt say where they benchmarked it, you cannot simply compare benchmarks, as they may have been more or less demanding.

[marldorthegreat]

Now if only AMD's low-midrange cards could follow suit and see similar improvements instead of being constant rebadges...

 

That would be just great.

Yeah, it would help if they had enough r and d budget to actually PAY for it.

[Ohlyver]

Yep, choosing settings to avoid getting bottlenecked by the 64 ROPs on the Fury, which, if you are playing on 4K you should be doing anyway. As it's a card meant for 4K, only light anti-aliasing if any at all is needed, unfortunately a lot of reviewers are using MSAA and shit on 4K which really is dumb as boots.

AA is somewhat dumb in itself when I have tried the NVidia DSR which allow for higher rendering res than the panel is. Tried a couple of games and man that was even better than using AA.

[sgloux3470]

They didnt say where they benchmarked it, you cannot simply compare benchmarks, as they may have been more or less demanding.

True enough.

[Nater]

So your retracting your statement completely.

 

No because there is still a better selection of g-sync monitors.

[Sam Z Man]

No because there is still a better selection of g-sync monitors.

In what ways? I would be interested in seeing the evidence that supports this claim.

[19_blackie_73]

forum is too green

I wouldn't call it to green, for me only nvidia cards suited my needs up to blender 2.74, only 2.75 which will come out soon will change the thing: now i can use also amd cards for cycles rendering.

If you're looking for a efficient card, you won't find anything except maybe the fury x on amd's side: 380 uses 70 W more power than a 960 and even 40W more than a 970; the old r9 2xx series had greater power consumption and went extremely hot - that are reasons i think many people will agree you can't ignore although they are (most times) cheaper and may have greater price to performance ratio. AMD makes it really hard to love them and recommend them.

BUT: I know we need competing companies to get new technologies and better prices so you need to have amd in the competition! I think the fury x is a step towards the right direction, but they should have done also a whole new generation in the 300 series based on a cutdown chip of the fury x like 980, 970 and 960 (but not with that crappy memory 500mb like space on the 970) which are efficient and smashes nvidia cards