I refer to the hardware. In which case, the Wii U is demonstrably far more powerful than the current generation. As it should be.
Will that translate to seeing an enormously obvious leap in terms of what we see on screen? No. but it will still be quite a bit better and obviously so.
"Moore's 'law'" was more a theory than anything else, though it played out very well for a long time. It should have been "Moore's observed trend" instead. LOL
and it relied on people working on chips to make it happen. Obviously the fallibility of man combined with varying talent levels, resources, economies, market, etc. would see to it that it would not be a perfect law.
Still, from the technology at hand now, the Wii U is far more powerful in every way. Will it be reveal in a leap forward the likes of which were seen by the 360 vs original xbox? Doubtful. But I wouldn't put it past GCN from N64 type of leap considering the hardware. Sure, the GamePad has to cost the hardware while it's doing its thing (despite what AMD would say), but there are ways for creative types to go full throttle on the graphics with minimal gamepad parasitic drag.
In other words, I agree with where you are coming from, but at the same time, I believe "far more powerful" to be accurate, though by far, I mean from California to Texas. Not from California to Israel, if you get my drift.
Really interested to see what the Crytek folks are doing with Wii U. If there is an engine looking for hardware to let loose on, that's it.
Moores law relied on a dennards cpu scaling, which is real and tangible.
in 1974 dennard postulated that MOSFETs continue to function as voltage-controlled switches while all key figures of merit such as layout density, operating speed, and energy efficiency improve – provided geometric dimensions, voltages, and doping concentrations are consistently scaled to maintain the same electric field.
This is where moores law comes up with tra.sistor density doubling every 2 years. It was his observation of dennard scaling with the advanement of manufacturing capabilities.
This is whefe the large majority of progress from 1974-2003/4 came from. Simply increasing transistor density and decreasing size, allowing for increased clock speeds as a nice boon too.
Moore estimated that his observation would only be sustainable at the rate he saw for 10 years. It went on for 30.
Everything practically in power grew linearly with time.... It was like a free lunch buffet for power increases.
Until we got into the early 2000.
We started hitting the limit of physics, the designs were too small and began leaking current into the substrate, and we were no longer able to mantain the same eletric field.... We couldn't double the transistor count in the same amount of time anymore... and soon we wouldn't be able to at all.
the buffet closed.
That's when other existing but largely ignored technologies began to be brought to the fore front.
First up was a concept called concurrency.
You already know what concurrency is, just not the old unmarketed name for it. Concurrency is splitting the software into seperate lines of work to be processed in parallel. Software multi threading. Unfortunately, not all tasks can be processed in parallel, so the gains received were rather small. Around 15% increases in performance. Then came chips DESIGNED for concurrency, with seperate threads in the hardware, this was marketed most known as hyper threading. Depending on how well optimized the software was, hains were generally logged from 15-40%. Very nice, but a far cry from the 2x gains of the free lunch era.
Then came processor concurrency, multicore, which gave much better boosts in gains. For a while, we were seeing progress close to what it was before...
Unfortunately, that was destined to be short lived.... For each core added the gains become less and less due to overhead like core and thread coherency, and amdahls law of parallel processing.
A 4 core at 2Ghz per core is nowhere near as powerful as a single core processor @8GHz (with the transistor count of all 4 cores).
Engineers are constantly, desperately trying to find new wells to dip into, now they're trying 'dark silicon', the problem is, each well is shallower than the last, and inherently more expensive... And all are little more than a drop compared to the heyday of dennard scaling.
Engineers are working harder and harder to scrape up smaller and smaller gains in power.
Silicon on insulator technology has pretty much hit its physical limit. Advancements will be considerably slower until a new material technology is found that allows dennard scaling, or something similar to kick advancement into high gear again.
That reality applies to game consoles too.
Its just not in the cards to advance games consoles this time like we've been doing for the past 30 years.
