Nintendo Switch Technical discussion [SOC = Tegra X1]

Xenoverse 2 is 1080p/60fps 100% of the time on XB1.

Xenoverse 2 is 900p/30fps (docked) 99% of the time on Switch + downgraded graphics.

Is it a comparable compromise to that of Snake Pass ? Basically 900p vs 720p + downgraded graphics.

Clearly not. The obvious conclusion is that Snake Pass is very unimpressive on XB1/PS4. This is why the gap is so small compared to the Switch. It is not complicated to understand.

Even if the developers used double packed math on Switch (very unlikely imo) the small difference still can't be explained compared to other games.

Also : 720p/60fps > 1080p/30fps.

 

The difference between Xenoverse on PS4/X1 to Switch seem rather similar to that of Snake Pass. Seems rather simple to understand for sure. 900p 60fps on Switch with two players on screen to be clear.

 

Which is 1% of the game... 99% of the time you don't have 2 players on screen... and if performances were uncapped on XB1/PS4, the gap would not change with 2 players on screen (likely 130fps vs 65fps).

 

I haven't played it yet, but I find this hard to believe. Not sure how you can control two players at once any more than you could in Tekken Tag. I plan to pick it up, so I will know for myself within a couple weeks.

@Recop

Good clean debating tonight. I enjoyed it. Good night.

 

You don't control them. They are controlled by the CPU (story mode).

But indeed, if you only select the 1vs1 mode (to play with friends or online) then you have a 900p/60fps game. But it's not a really fair comparison...

 

Let's get to the topic for a little bit. On neogaf some people said what Switch GPU have some features what XB1/PS4 don't have. But there wasn't some details about it. Can anyone tell more about those features?

 

I guess they're talking about FP16 ? But other than that... (and it's present on PS4pro AFAIK)

 

Could also be the tile based rendering, which can increase performance a little bit and saves power, while also reducing the pressure on memory bandwidth.

 

We've already talked about double-rate FP16. There is no magic sauce in Switch.

 

But it is true that the Switch could get a performance boost with double FP16 though. Up to 5,10, 20% ? That is the question...

 

Switch games are already making full use of its technical features. They have to because its so low clocked and sips low power when in portable mode. If they didn't, the games would perform even worse than we've seen.

 

I'm talking not about magic sauce, I'm talking about real tech things.

 

Exactly. We've already talked about everything because its been known items since its using 2 year old Nvidia Shield products but clocked a lot lower.

 

I've missed it, because I wasn't interested in Tegra at all before Switch.

 

New info about resolution. https://mynintendonews.com/2017/09/23/doom-for-nintendo-switch-docked-and-undocked-will-run-at-720p/
So there is probably chance for some graphics improvements in docked mode.

 

If that is all you can answer, why answer at all? This a technical discussion (about Switch even, not Tegra X1 per se), that was a technical question. So, if you know all the answers you might as well give him some? There are bound to be more things that TX1 has that PS4 SoC does not, apart from FP16. The tile based renderer as an example, which was one of the things that helped nVIDIA get a generational jump in Maxwell without a jump in fabrication node.

But if you think we have already talked about everything I guess we might as well close the thread lol

 

Older GCN in PS4 and XBO didn't have delta color compression either.

 

When someone comes around and keeps asking the same questions in the same way that seems to indicate they're looking for some sort of magic-sauce it gets tiresome. I'm just cutting off the craziness before it starts here again. It took a long time to stop the insanity of magic-sauce believers in the Xbox/Wii/WiiU threads on this forum, so we're being more vigilante now.

The existing Tegra threads are full of info if anyone truly wants to know.

 

I understand but the Tegra thread is about Tegra alone. It doesn't make direct comparisons with PS4 and Xbox One GPUs there. Which was what the question was about.

 

And that's 6+ Watts measured via the USB line from a charger, where the system is still likely to be seeing additional draw that won't be present in battery only mode.

The Switch will probably run on less than that when powered only by internal battery (or batteries, including controllers).

By my possibly wrong-headed calculation, 3.7v x 4.31 Ah / 3h = 5.3W for 3 hours gaming.

2.5 hours battery life would be from ~6.4W draw, 3.5 hours would be ~4.6W.

So about 2~3W from the SoC in mobile mode. Maybe 3~4W if you're really pushing it in a battery devouring game.