Yeah, I thought the 750TI was more powerful before I looked at the site I linked, I'm gonna agree with the site you posted and say it is more powerful.

Regardless the OP is kind of silly. Sure bring to attention that TFLOPS != Performance on it's own but don't do something just as bad. I mean using two Desktop GPU's TFlops to compare the potential difference between a MOBILE GPU's and a Console GPU's TFlops....

The Ipad Air 2 has a 27 watt/hr battery
if the graphics chip used 90 watts you'ld get 15 minutes use before you had to recharge it, Is this what you want the NX portable to be? Only usable for 15 minutes! [smile]

A Tegra X1 only uses 21watts to do around 512 GFlops though, not 90watts. Going by what people say the Nvidia Shield Console uses.

Nintendo would just need to use a really big battery.

What?

Shhh.  The nerds are nerding and it's beautiful.  Now let's each take a shot each time we read a post we can't understand in this thread.

I think the math  works out more like 18 minuets... but hes not wrong.

a battery of 27watt/hr, with a 90watt power consumption when gameing, would only give like 18minuets of gameplay.

27 watt/hours x 60 minuets/hour  / 90watts = 18 minuets.

That means a Tegra X1 with a battery like that would give around: 1 hour & 17 minuets.

Solely because it's clocked higher. I was wrong about the 7850 running at 1GHz (mine does but it was an overclocked card),  the default is 860Mhz. Radeon 370 runs at 975Mhz, or ~13% faster. 1.76 Gigaflops x 1.13 = 1.98 Gigaflops.  Again,  they are the same architecture,  yet a 13% clock increase has erased a 30% speed deficit by your numbers, demonstrating that it is more recent drivers optimized for the later games used in your comparison that makes the difference. 

What you have proved is that flops is  silly metric to use to compare performance with unless you are  only doing compute.  But that's all you can really conclude from your data,  nothing more. 

if NX is really going for Nvidia Tegra, I wish it will use Tegra X2, Pascal architech, rather than Maxwell.

OP's comparison doesn't really matter since NX is DOA power wise if the rumors are true ...

That thing practically has no hope of ever surpassing current gen HD twins or getting parity within the X1 ...

Even the Pascal architecture won't save this thing from being slaughtered by 100+ watt devices performance wise cause that thing has to operate within a strict power budget ...

Personally speaking though I'm not all that interested about the compute power of the NX, what I'm more interested are in the hardware features like conservative rasterization or raster ordered views but I would've been way more interested if they went with Intel to see what they could do with that nice gfx IP providing that their advanced technology group was up to speed on the state of the art ... 

That was my intention all along. You seem to take "my data" too seriously, I even put on the OP that I was too lazy to calculate a real average avantage for the 950, so I invented one instead. How come that didn't raise any flags about how serious I was being in my calculations?