Look, from AMD's perspective, $550 is a good deal for the HD 7970 and they are shocked that anybody but a Nvidia fan would say otherwise

Let's look at what we really have - 25% more performance for 10% more dollars than the GTX580.

That is unprecedented. Terrible for a NEXT-GEN part on a full node lower process. It does not bode well for the HD 7800 series pricing wise and it could be a giant back-fire to AMD who is used to spoiling their fans.

And the performance bar is now lowered for Nvidia. They can go 25% more performance for 10% more dollars than the HD 7970. It may be good in the short-term for AMD, but bad in the long run - it's going to depend on their respin of the HD 7970.

Well. There will be a re-spin. I'm positive of that.

i said it first - 2 days before the HD 7970 launched:

http://alienbabeltech.com/main/amds-upc ... video-card

Just look at the title:
AMD’s Upcoming HD 7970 Exposed – a Short-lived Video card?


My choice of a title may have *appeared* to have been sensational; as it turned out, it is true




6 months to the HD 7980 . . . and counting

Probably 4gb.

Huh?

The card will probably be 4gb. the 7980. LOL.

vRAM can be 1.5G or 3GB for the 384-bit interface without using mixed RAM as Nvidia did on some of their lower end cards

-- next stop would be 4.5GB or 6GB
-- about 2016 as far as i can tell

There goes that.

No doubt AMD is late in that light. They are late by two yrs if they were trying all along but they wasnt. They were on a different path and considered Nvidia's "the wrong direction". AMD were slow to see the light. Its obvious Nvidia took the leap first, they lead the way. But AMD shamefully was wrong. They are goin the fermi path, following nvidia's lead and by this they are late to the game.

What i was getting at is: although they were late to pursue, their transition wasnt terrible at all. Their execution is impressive to me. It wasnt smooth, and i am surprised they got it together so quick. They got a product to sell in very short order. In this i commend them.

BUT-

It will become apparent how much AMD is behind once nvidia drops kepler. There is no way around this. AMD dragged their feet too long and nvidia is yrs ahead in this GPU/compute path. The 7970 will not be enough to stand up against kepler.

For now, AMD has the crown. For now they have a decent product that competes well with nvidias gf110. This is what they have. Their 7950 will compete under the 580 and their 6970 is a notch above it. Kepler will be out of reach. But what can you really expect when AMD sit of the fence and waited sooo long before they jumped in.

That said, I think AMD has gained some ground impressively fast. Of course its not gonna be enough. But instead of 2yrs behind like they should be, when the smoke clears they have gained ground. They have gained a lot of ground real fast.

It's interesting that the 7970 seems to really shine at 2560x1600 vs the GTX 580 (probably due to stellar memory bandwidth) and at eyefinity resolutions with FXAA. It's a bit unexpected to see so many review sites using FXAA in their benchies this soon, already! But when 8x MSAA is used, the performance plummets to levels closer to the 580 - probably due to only 32 ROPs.

What I'm a bit disappointed in is that FP64 performance remains at only 1/4 the FP16 performance, like that of the VLIW4 cards from AMD. I was kinda expecting GCN to do it at least 1:2 ratio or better, like the Fermi-based "unlocked" Tesla cards from NV.

Have anybody done approximations on the 4.3 billion trannies with all of the statistics to logically deduce if parts of the 28nm chip could have been disabled? A 384-bit bus shouldn't take that much die space at all, plus there is still only 32 ROPs. The increase of shader count only went up by 33% but I'd assume that the GCN arch takes about 15-20% more trannies per shader (with the resulting performance boost of around 10% per shader). Who knows how much more trannies are needed for the new DX11.1 features? The doubling of L2 cache does seem to take up some space, but heck it's still ONLY a measly 768K (compare that against 12+MB on the CPUs that take up far less trannies). 2048 shaders and 128 TMUs seem like a fully-rounded number, so it's a bit hard for me to imagine the unlocked number of shaders if there were more... perhaps it's the ROPs (really being 48, advertised as only 32) since they take up a LARGE amount of die space/trannies? I'm afraid that AMD could be playing us once again like they did with the # of Bulldozer trannies (first announcing 2B, then a few months after the launch admitting the "error" and correcting it down to 1.2B). Especially as the HD 5830 and 6790 show some "impossible" stats, along with the Barts series claiming VLIW5 when they behave like VLIW4, I've taken a healthy dose of skepticism this time around.

Perhaps behardware.com would be doing some synthetic tests on the bandwidth and pixel fillrate capacity to see if the specs hold true (like they did on the 5830 and then the 6790 http://www.behardware.com/articles/827- ... -6790.html )....


Maybe we'll be seeing an unlocked 48 ROP part soon that allows the 384-bit bandwidth to be used to its fullest?? My guts say no.. that we'd merely be seeing a better-binned process like the RV790XT (4890) refresh of RV770XT.