lostinblue
Banned
That's a (very) dumb way to look at it. Xbox 360 is 240 GFlops, full stop. CPU Gflops are not used for graphics therefore it's just overhead for something else. AI, Physics, calculations you name it. Adding them up it's simply dumb as fuck as it's never that linear.Assuming the 360's lower end of FLOP output holds true, we are seeing AT LEAST 300 GFLOPS, and 4 time that is 1.2TFLOPS... which the Wii U is not...
Sure Wii U software might be looking to move some of that to GPGPU, for the sound processor and other dedicated parts to take part of the load off or perhaps some of those these days belong on GPU anyway, but *perhaps* most code will instead be optimized to run on the CPU in general purpose, who knows. As said, totally not linear; most GFlops on the GPU will in the end be used solely for graphics though, that's a given.
Also, most people here are forgetting something, if you want a 720p X360 game running "as is" at 1080p you need 2.5 times the fillrate/floating point performance; that amounts to going from 240 GFlops to 600 Gflops. Resolution and pushing pixels is the single thing whose hit hasn't been softened by acelerations, optimizations and the like, it simply costs that much and requires a order of magnitude to improve whilst doing the very same thing merely pushing more pixels.
On Wii U's case, assuming it's a 372 GFlop part and comparable that way, then PS4 and X720 need to have 930 Teraflops to do at 1080p whatever Wii U is doing at 720p. Of course they surpass that, but X720 is going to be 1.2 Teraflops, which is not a huge overhead providing it's doing 1080p (if it is doing 720p though, it's very palpable of course).
I'm not commending Nintendo on the system final specs, make no mistake, playing it low is a risk, and it's certainly not doing them any favours, nor am I saying it'll benefit from multiplatform across the whole generation, but I also believe that's probably due to ill developer intentions and bad preconceptions, as most of the devs dissing the system haven't bothered to take a dip on it, they probably looked at specs and concluded they'd be bound by the same bottlenecks the current generation has, while we know Nintendo is all about removing them. Of course Nintendo is still to blame for putting themselves on the situation where support depends of developer goodwill; but it is nevertheless very doable, unlike GC to Wii they moved the GPU well past 2005/2006 tech (the current generation starting point).
Read above.Anyway, is he right to think Wii U is closer to PS4/Durango than Wii was to PS360?
I think he is technically right, but I was wondering, does it matter? The difference still seems very large.
As for the rest of the equation, really depends on what the basis for next gen development become. See, if you do a game with tesselation geometry in mind then the Wii version and PS4/X720 might have the same assets with varying LOD's. That's very different from past generations when downports had to be optimized and lower polygon geometry had to be created and optimized.

There's palpable seeing the wireframe difference, yes, but with good tetxuring perhaps not so, plus providing it runs in a fluid manner across all systems it's unconsequencial; just like whilst running a game on a computer across various configurations, sans the lack of optimization typical of PC software.
Of course, most developers have no experience using this tech and would prefer to use it as a means to spam insane LOD's over an already overdetailed 3D model rather than making a game to be barebones on it's knees so that it can run across all systems easily, but the later is really the best way to use it as the end result would be pretty much the same on the higher end system (using tesselation as a core technique rather than a plus). Also note, rendering tech and tools are becoming more and more scaleable so it only makes sense that games follow that tendency if developers can get past their pride of "being focused on the higher end" and the "we look forward not backward" bullshit line of thought; coding close to the silicon of a high end machine (or any machine) can be considered nuts by now; instead code and assets must strive to be abstract enough.
Also, perhaps the Wii U is to X360 "the same Wii was to GC" on a order of magnitude; but you gotta consider the rubberbanding of the whole scenario; PS2 was 39 times less powerful than X360 is, GC was 28 times less powerful (this doing only GPU versus GPU math), and the Wii was 18 times less powerful than a X360 (not to mention lacking in feature parity). Wii U will be at most 5 times less powerful when compared to PS4 (with X720 staying somewhere in the middle), it's not the same situation.
Somewhere, sometime not supporting a platform because of such power differential won't be justifiable anymore, be it because it's already powerful enough or because also selling said software on it is a worthwhile option to get the whole investment back. That time could be now, but developers often behave like juvenile indie-wannabe elitist clusterfucks so the jury is out. One thing is for sure, the market probably won't grow this generation and development costs for PS4/X720 are supposed to increase, so making it a two console environment all over again might not be wise with a huge chunk of games not providing an experience that couldn't be achieved on current gen. And one could pull a lot more than what's being pulled out of current iOS/Android mobile devices, but it's not worth the investment, on a similar line of thinking, one could pull more out of current gen and will always be able to pull more out of upcoming next generation systems for they're very complex systems at this point; but is it worth to? The answer, paired with industry risk, cases of flop, losses and layoff shenanigans that are common now is increasingly no.
What I'm saying though is that Wii U having some features the past generation lacks (or lacks in a usable way) has the potential of single handedly making it a whole different system when it comes to capabilities. Tesselation in theory makes it so that you could pull the best X360 closed scenario graphics in a open world setting keeping the 240 GFlop spec, because it can keep the further away geometry simple (keeping the end result the same) sparing you loads of polygons and coordinate tracking or complex geometry swaps. You can effectively do more with less if you tackle it, but at the same time it's not like going to the code and adding tesselation=yes.
No one knows how to use it for that effect still.