It is very unlikely that any modern day Android game will be pegging all four A15 CPU cores (let alone pegging two in fact). So in most games, GPU utilization % will be much higher than CPU utilization %. And with the vast majority of Android games, Kepler.M will not come close to being fully utilized either. Last but not least, the R3 variant of Cortex A15 on 28nm HPM used in Tegra K1 has superior power efficiency compared to the Cortex A15 variant in Tegra 4.
Once again, 5w is the TDP for the entire SoC. For CPU-intensive apps, most of the power is allocated to the CPU (and vice-versa with the GPU). Any scenario where both CPU and GPU are pegged at the same time is pretty unrealistic. It would also be pretty unrealistic to expect Kepler.M to have the same clock operating frequencies in a device like Shield compared to a tablet or phone.
No Android game even remotely comes close to stressing Adreno 320, and that is not going to change when K1 ships. The same "cater to hardware fragmentation" syndrome that happens for PC gaming is also now happening for mobile.
IMO Nvidia should just stop persuading Android mobile makers to use their chip like Intel and come out with their "consolized" phone/tablet with great controls and games as a killer app.
Maybe hack a customised version of Android with Steam/even just port SteamOS over entire to get the taxing enough/non mobile focused games?!
They'd still need to get people to recompile/test them for their arm chips of course. Maybe not too bad, cf that SSamIII demo of theirs, if the graphics engines are running basically unchanged.
Looks impressive when plugged into the wall, in a chassis big enough to fit a quad core i7 in it. Let's see how it looks in a fanless tablet...
I don't see how is that impressive either compared to a now shipping SoC on a $350 phone.
That was kind of my point, though I phrased it weirdly. Tom's Hardware compared a part in a desktop AIO, with desktop cooling and desktop power supply, to tablets and phones...
I don't see anything outstanding here for a late 2014 product, in fact 48 fps is slower than expected and slower than Nvidia's 60 fps estimate. Maybe K1 in this system run slower here. 3dmark Graphics score looks even worse.
That was kind of my point, though I phrased it weirdly. Tom's Hardware compared a part in a desktop AIO, with desktop cooling and desktop power supply, to tablets and phones...
didn't notice the GFX bench where K1 is 48 vs Qualcomm S800 23 (108% ) faster, now if normal android games is more like 108% faster than this is a critical advantage, any reason why GFX bench gets such a larger performance uplift compared to 3d Mark?
Could be wrong but... Doesn't the physics benchmark scale with cores, and S800 is a QB and K1 is a dual?
Don't forget that this prototype monitor already has significantly reduced frequencies for CPU and GPU vs. what we will see in a typical TK1 tablet (even though there is no need for that in this form factor). It is a prototype after all.
Here is comparision between a 5" smartphone and a 21" AIO desktop with Tegra 4:
http://gfxbench.com/compare.jsp?D1=HP+Slate+21&os1=Android&api1=gl&D2=Xiaomi+MI+3+(Tegra+4)&cols=2
The GPU clock is unknown, saying it's significantly reduced from what it'll be when shipping is an assumption.
Maybe hack a customised version of Android with Steam/even just port SteamOS over entire to get the taxing enough/non mobile focused games?!
They'd still need to get people to recompile/test them for their arm chips of course. Maybe not too bad, cf that SSamIII demo of theirs, if the graphics engines are running basically unchanged.
It's going to be hard enough getting developers to port to SteamOS, never mind SteamOS on ARM...
More benchies this time from the note 7 inch reference tablet.