so it can play tetris?

I have a GTX 275 and it still runs everything of the new games just fine. Of course I don't play at resolutions like 1080p, but 900p still looks fantastic (My screen doesn't do more anyways :P). I'd rather play on 900p with AA than 1080p without AA + AF. People are too spoiled with resolutions like 1600p. 

How much better than a Radeon HD 6990?

Irrelevant to the question in hand though, since it would be nearly impossible to quantify where both architectures stand right now. Anyways even if you take into account only the RSX verse the GTX 580, the difference in raw number crunching would ammount to some 6.9 times, no matter how crap the RSX is I doubt you would see a three times increase in efficiency alone! The difference wouldn't increase too much even if one was to take into account, say, double precision operations, so...

There are major 3 elements which determine the performance of a GPU.

1. Shader throughput  - your GFLOPS data

2. rasterization rate - # of triangles per second it can draw.

3. Memory bandwith 

You're only talking about shader throughput which is only 1 factor and really doesn't matter without knowledge of other 2 elements.

For example, GTX580 has 1/2 the shader throughput as Radeon 5870, but its probably 1.5X as powerful due to higher memory bandwith and 3X the rasterization rate.

I'm too lazy to find all of these RSX numbers and compare them to the GTX 580, but if you feel like being proven wrong, look them up. GTX 580 has roughly 20x greater overall performance...

Come on disolitude, framerates are not going to be 50x better even considering all that.

Edit: you said 50x. Now 20x? You're going to back down to 6.9x?

GTX690 = X50 theoretical performance over current gen consoles

GTX580 = X20

Only reason why I brought 580 in to the mix is because a while back I read that the GTX580 has 20X theoretical performance of current gen consoles.

GTX690 = GTX580 X2.5 (roughly the performance increase over GTX580)

x = 20 X 2.5

x = 50

No change in my statement is necessary. :)

Edit - if well optimized game running on PCs with an RSX like GPU and a GTX 690, and exactly equal settings, and CPU/memory bandwith, RAM wasn't a bottleneck factor...you most likely would see a 50X getter frame rate on the GTX 690. 

Resident Evil 5 for example...I've personally benchmaked it and single GTX 580 gets 200+ fps on 720p with CPU being the bottleneck. I will bet you any money that RSX like GPU would muster 5-10 fps at same settings...

On paper the GTX580 is arround 10 RSX's welded together but with Unified shaders, Tesselation, Compute functionality etc which would crush the the RSX utterly in real world performance due to reduced bottlenecks and new hardware functions that would allow the 580 to do things that aren't even possible on the RSX (tesselation, compute functions like Physics accelleration, decompression of assets, advanced SSAO, AA and GI effects, and even supports working with better compression of assets without decompression so textures etc take up less working memory and less time to load). There is really no way to accurately compare the two.

And benchmarks are here. The card does exactly as it on thinks it will... Only thing that surprised me with these benchmakrs is how well the 7970 scales in CF. Single 680 usually beats 7970 but on 3 screens and in SLI crossfire, 7970 seems to trade blows. 

Read - AnandTech
Read - HotHardware
Read - Legit Reviews
Read - PC Perspective

That's an awesome card, but it will be no where near any next-gen hardware. Too new, too expensive, too much heat, too low yield, etc.

When I think about the price of just this card, its crazy to me that people buy these so early. I guess that can be said for any new tech though. There's always a person that sees value at some price point.