K1 compared to a typical Desktop Kepler GPU

[ams23]

The graphics performance makes a big leap every year. And Tim Sweeney was on board with UE3 last time (at least on iOS and Windows RT).

A perf. leap is only half the equation, because perf. still needs to be high enough to be playable 🙂

Kepler.M inside Tegra K1 is arguably the first ultra-mobile GPU that has comparable (or better) overall GPU + CPU performance compared to PS3 and Xbox 360

 

[NTMBK]

A perf. leap is only half the equation, because perf. still needs to be high enough to be playable 🙂

Kepler.M inside Tegra K1 is arguably the first ultra-mobile GPU that has comparable (or better) overall GPU + CPU performance compared to PS3 and Xbox 360

Oh man, this is just funny. Go watch that video I linked, listen to what Tim Sweeney says right at the start.

...whereas this brings mobile up to current generation high end console status, basically.

This was talking about Tegra 2, 4 years ago. None of this hype is new. NVidia have taken us around this merry go round before.

 

[s44]

Basically its previous generation console like tech, which makes it performance wise around 2005.

It has significantly more shader power than the 8800GT, which was state of the art in 2007-2008.

 

[ams23]

This was talking about Tegra 2, 4 years ago. None of this hype is new. NVidia have taken us around this merry go round before.

Think logically about this. Tegra 2 may have had DX9 API support and a proper DX9 architecture, but it was way too slow to play most DX9 console games. Obviously Sweeney was optimistic at that time, but the reality of the situation was that perf. was still too low at that time. And at that time, Android was the new kid on the block too.

Once it becomes feasible (from a performance perspective) and monetizeable (from a financial perspective), high quality games will come to Android. It is not a question of "if" but "when".

 

[NTMBK]

Think logically about this. Tegra 2 may have had DX9 API support and a proper DX9 architecture, but it was way too slow to play most DX9 console games. Obviously Sweeney was optimistic at that time, but the reality of the situation was that perf. was still too low at that time. And at that time, Android was the new kid on the block too.

Once it becomes feasible (from a performance perspective) and monetizeable (from a financial perspective), high quality games will come to Android. It is not a question of "if" but "when".

I would kind of agree with that, actually. It needs both components; technically feasible and financially feasible. I have no doubt that parts will get there technically; but I don't see the business case working. Not until a big, big player (Google, or Amazon, or someone of their size) throws their weight behind an Android gaming platform. If Amazon come out with an Android "console" with their own marketplace for it and massive support, then we could very well see a boom in "AAA" Android games. But in its current state it's just not going to happen. Ouya tried and failed, Shield tried and failed (though the game streaming tech is cool).

It's a chicken and egg problem; developers won't make expensive AAA games without a huge install base, and users won't buy Android gaming boxes without AAA games to play on them. This cycle is normally broken by console developers putting money down to develop first party games, and to secure exclusives for the platform. Without someone "owning" the platform in that way, and drive both game development and system sales, it won't take off.

 

[ams23]

Yes, it may take a company like Google to help kickstart Android gaming (just as they helped to kickstart Android tablets with the Nexus 7). One way or another, NVIDIA needs to play a big role in helping that come to fruition.