Earlier this week, Bloomberg exploded Japan’s report on a rumored Nintendo Switch “Pro” version with a powerful claim: all these rumors about a “4K” switch may, after all, be true. The latest report on Tuesday teased a vague bump in specifications such as clock speed and memory, which could make the switch run better … but jumping all the way to 4K resolution will need a massive boost from the 2016 system’s current specifications.
What made the report so interesting was that it had a technical answer to the seemingly impossible rendering challenge. Nvidia, Nintendo’s exclusive SoC provider for existing Switch models, will remain on board for this updated model, Bloomberg said, and this contribution will include the exciting, Nvidia-exclusive “upscaling” technology known as Deep Learning Super Sampling (DLSS).
Since the report was published, I’ve been thinking a bit, and I can ‘t shake a certain emotion. Nvidia has a much larger plan for the future of average users’ data processing than they have publicly given.
Makes games smoother – no doubt easier than Tom Cruise’s face
DLSS, for the uninitiated, is a system that has been developed for a long time and relies on thousands of hours of existing video recordings, which are largely taken from 3D games, and which are run through a machine learning model. When you hear about “AI-trained” systems such as playing video games or generating crazy names, they usually work in the same way to reliably detect successful patterns. They study and duplicate existing data (dictionaries, Twitter conversations, Tom Cruise’s face, videos of StarCraft competitions), then they go through double, triple and hexagon check runs to confirm how well a computer can organically recreate the data before losing their systems in real world tests.
However, these tests are usually run with a supercomputer that manages this data. This is where DLSS fluctuates sharply. As a consumer quality offering, DLSS requires Nvidia’s “RTX” graphics card which (theoretically) costs as little as $ 329 – and nothing else. No additional CPU, RAM or other peripherals required. The idea is that RTX cards come with a plate of “tensor cores” baked on the silicon, and these are devoted to the mathematical groundwork with tasks such as image interpretation and translation.
This works in part because DLSS is focused on a relatively non-volatile prediction scenario: what a lower resolution image would look like if it had more pixel depth. The shape of diagonal lines, the leaves of swaying tree branches, and even the letters and words on a street sign are undoubtedly easier to predict – at least when driven by a base with lower pixels – than a StarCraft the player’s reaction to a Zerg rush. While DLSS ‘first launch left a few things to be desired, the system has matured to a point where moving 3D images with DLSS processing generally look better than their higher resolution counterparts, temporal anti-aliasing (TAA).
Side-by-side comparison of DLSS with lower resolution and TAA with higher resolution shows strengths and weaknesses, but as of 2021 it has become a wash. This sink has significantly higher frame rate and lower processing requirements for the DLSS page, since the traditional GPU primer is not only pixel-bound, but also texture-bound. DLSS ‘knowledgeable upscaling not only reduces the need for clock speed and processor cores; it also translates textures with lower resolution with more fidelity, so you can no doubt skip, skip and skip high VRAM requirements.
Calling BS on another BS satellite view
Nintendo Switch does not have an exclusive license for lower amounts of VRAM, lower clock speeds and rendering scenarios with limited shape sizes and thermal envelopes to work with. Where else can everyone be helpful? Hmm.
Nintendo Switch is an ingenious platform as a starting point, especially since we have said the same thing over and over again: “% GAMENAME% is great to play in Switch’s portable mode, where it comes close to 720p resolution, but blows the same action on 1080p TV. one of yours looks like rubbish. “You, the skilled player, should not seriously Consider using your switch for video games, despite statistics showing many people wanting a good video game experience (and an ever-growing market where more people than ever are in a hurry to buy Xbox, PlayStations and switches for their TVs ).
We are 22 years past Nintendo’s last step towards that concept.
A DLSS-equipped “Switch Pro” (not a final name), as described by Bloomberg, would achieve resolutions of up to 4K when docked and connected to a TV. Assuming that DLSS in this scenario requires dedicated tensor cores, we are currently wondering where Nvidia will beat said cores. Are we looking at a “shared motherboard” scenario, as we have seen on both the Xbox Series X and PlayStation 5, where the TV-plug dock takes on certain heat-generating elements like a block of tensor cores? Or will Switch Pro’s basic hardware contain tensor cores on its own SoC?
My completely speculative guess is the latter for a couple of reasons. First, rumors of higher CPU clocks and more RAM on the base unit suggest that Nintendo is planning to roll out a brand new system, as opposed to additional hardware that will increase the base. Nintendo has experience delivering additional hardware to home consoles in Japan – such as the Famicom Disk System (which added more power and extra audio channels), BS Satellaview (which added RAM) and 64DD (which added a co-processor.).
But we’re 22 years past Nintendo’s last step towards this concept, and after the mass confusion caused by the Wii U, I do not see today’s Nintendo wading into the consumer confusion of questions like, ‘can I put an older switch into a Switch Pro dock for more? power? “