The Mobile chips have become 75% more efficient in the last two years. To a significant degree due to going down to 16nm, but it's still notable. They will hit a wall eventually (or slow down) but for now, there seems to be steady progress in mobile power.

Even taking into account a RISC architecture much nimbler than x86, it'd need a giant leap in power efficiency meant as performance per watt. Also, GPU is the real power hog, CPU isn't anymore, and mobile GPUs didn't find some secret graphics holy grail PCs and consoles cannot have, they can only count on the usual efficiency tweaks also portable PC and PC and console APU GPUs use, and the latter two still can't reach desktop GPUs performances anyway if they have to keep consumption and heat under control, for laptops, and costs for consoles, so the ony exclusive tweak mobile GPU can count on is having to generate a much lower number of pixels and polygons to get on small screens the same fine grain a desktop GPU have to generate on a larger one (yes, there are also insanely hi-res small screens on hi-end smartphones, but don't expect anything but some videos to fully use them and even these rare cases will eat battery charge even if nowadays video decoding has become very sober in power requirements compared to games HD 3D acceleration).
Anyway, if they meant they'll exceed, on small screens, the PERCEIVED quality compared to what 8th gen home consoles can display on large screens, this is not unlikely, but they won't surely outperform them as brute computing power. Use those chips to generate images for a large screen and their grain and quality won't look so fine anymore.
This said, nice to see ARM avenging RISC architecture that Intel tried so many times and for so many years to kill.

LMAO

Yer rite

off topic what does this mean

(Warning: Division by zero in /data/r5/nfs/ioi/gamrconnect.vgchartz.com/my-topics.php on line 177)

Lol nice try ...

When i mentioned gigaflops I meant single-precision calculations not the half-precision crap that mobile device manufacturers use in their skewed benchmarks.

Guess you don't know the difference between half, single, and double precision floating point operations. They are all exponentially harder to perform than each other. Half precision calculations are basically useless to anything related to graphics.

The areno 430 is known to throttle severely for any task that lasts longer than a few milliseconds. Actual real world benchmarks peg this GPU at 50-60 single-precision gigaflops.

Shame the devices would overheat within about 1minute of use.

I have big doubts. The architecture of the CPU and GPU is not even close to what we have for the PC.

Frankly, I should hope so. The R/D process for the PS4/XB1 started years before their launches in 2013, and they weren't based upon the highest end tech at the time (CPU especially). I would expect hardware launched in 2017 would exceed them in power.

by then, we would have been 4 years into next gen. Mobile will never catch up to console