fatslob-:O said:
fallen said:
fatslob-:O said:
fallen said:
ViktorBKK said:
F0X said:
ViktorBKK said:
The Wii U has more ram, thats it. Everywhere else it is lagging. You obviously do not understand the correlation between semiconductor lithography and power consumption.
Wii U : 40/45 nm & 40 Watts
PS360 Slim Versions: 40/45 nm & ~100 watts
Do the math.
|
Care to elaborate more? I'm led to believe that technology is supposed to become more effiecent over time.
|
Chip designs get more efficient over time, but engineers are not wizards. Especially when it comes to GPUs, where things are highly paralellized. You can improve a design and gain maybe 15% performance under the same thermal ceiling. You can't ever get 100%. Ever seen a 100W AMD card beat a 200W nVidia card? When both cards are in the same proccess node, it cannot happen. Also consider that the GPUs from all 3 current gen consoles, come from the same fab(TSMC).
|
not really true because smartphones are approaching ps3 power and they draw like a watt or two.
mind you, i believe the wii u is quite weak. just correcting you.
|
Not even close!
|
Not even close what?
Go look up the benches. Even something like Tegra 4 is as powerful as a 7800 GTX which is the GPU in PS3.
Let alone Project Logan, Nvidia's next mobile GPU, which they explicitely state is more powerful than PS3 (google Project Logan).
You may not see mobile games that look as good as PS360 yet, but that's for many reasons. The raw power is increasibgly there.
|
Project logan is gonna fail like the rest of the shitty tegra chipsets. 
How can they get more effieciency with the same node ?
|
Huh?
Yeah Tegra fails, but that doesn't matter. All the other competing chipsets are just as powerful it was just an example (also I was wrong and edited, Tegra 4 isn't as powerful as PS3 GPU, it's the newest Qaulcomm SOC)
They get more efficiency from the same node with better designs? I dont know, for whatever reason mobile performance increases extremely quickly.
Probably because it's relatively low to begin with, so there's more room to grow.
There's also a die size variable. A 400mm GPU on 28nm should be 2X as powerful as a 200mm one, all else equal. The max current size is around 550mm. But if you're below that you can increase power just by getting bigger (more transistors) on the same node. So if Qaulcomm makes a 28nm mobile SOC that's 100mm, and then the next iteration they make 150mm, then the next one 200, etc, they can grow that way without decreasing node, at least until they hit the 550 limit (and I dont think mobile is close to that)
