A new, custom Tegra heavy on machine learning and ray tracing.
In the two years that followed, a wealth of overwhelming evidence has essentially confirmed that they were right.
But let’s kick off with a basic, fundamental question.
Can we say for sure, with absolute 100 percent certainty that the T239 is indeed theSwitch 2chip?
That’s a negative.
For its own endeavours in the automotive and robotics market, the firm already has the T234.
After all, going back to the first Switch, no custom processor was involved at all.
For Switch 2, Nvidia has nothing in the parts bin that’s appropriate for a successor.
The latest Tegra processors - like the T234 - are built for an entirely different, more demanding market.
Let’s be clear here: the existing T234 is a monster of a chip.
With a die size of 455mm2, it dwarves the Xbox Series X processor at 360mm2.
The CPU cluster will be quite different than T234’s, for starters.
The difference here is that there’s enough bandwidth for full HDMI 2.1 support, assuming Nintendo uses it.
Other spec leaks derived from Nvidia’s Linux distribution are intriguing.
There’s one more component of T239 worthy of comment - the FDE.
It’s an entirely new hardware block not found in the T234.
FDE stands for File Decompression Engine.
There are a few question marks though.
What clock speeds should we expect from CPU and GPU?
How fast will the memory run?
This is a question that’s key to understanding performance limitations in a mobile processor.
There’s also the question of whether T239 has the deep learning accelerator from the T234 or not.
I believe this is crucial to a potential Switch 2 and its use of DLSS upscaling.
I’d be surprised if docked clocks exceeded 1GHz and it could be a fair degree lower.
1.0GHz also looks like the sweet spot for CPU clocks too.
However, these are calculations for a larger Tegra chip that’s definitely on the 8nm Samsung process.
What kind of performance should we expect from T239?
But there’s more: a smoking gun, if you will.
It was called NVM.
The Nvidia hack describes an NVN2 graphics API.
So, what should expect from the T239 in a mobile games machine?
This was achieved via a more modern GPU and more memory, for starters.
I’d venture to suggest that Switch 2 follows that trend - and goes further.
It’s that last component we’re looking at more closely here.
So why not simply get a 3050 laptop then?
There’s one sticking point though - the 2050 only comes with 4GB of RAM.
I’d expect to see 8GB or even 12GB of total system memory in Switch 2.
However, the fully enabledFortniteUE5 experience using Lumen, Nanite and Virtual Shadow Maps does work.
Still, it does run.
you might use hardware RT.
I also triedDeath Stranding, benchmarking its GPU-heavy intro.
At 1080p native, the sequence averages at 34.9fps, which does support the last-gen horsepower narrative.
At 720p, frame-rate rises to 52.5fps.
I’ve also run this bench at DLSS quality mode too - 67 percent of native res.
34.9fps average increases to 44.7fps.
DLSS support: useful for Switch 2 or absolutely crucial?
However, the higher the output resolution, the higher the cost of DLSS.
It’s also important to stress how important DLSS is to the whole Switch 2 enterprise.
DLSS 4K ultra performance mode upscaling from a native 720p?
It just needs to provide a decent image on a 4K living room TV.
There is still controversy and many unresolved questions.
Questions and concerns: is T239 viable for a handheld?
However, there are some more legitimate questions and concerns over the T239.
A larger chip requires meatier cooling which may not be a good match for a Nintendo hybrid handheld/home console.
Any modern game seems to run just fine on it, assuming you adjust controls accordingly.
And really, the magic is going to come from the developers themselves.
How big is it?
What kind of battery life is being promised?
What will the games look like?
To what extent can existing Switch hardware - including its JoyCons - work with its successor?
Could this be the most forward-looking hardware design from Nintendo since the GameCube?
