How Will be Switch 2 Performance Wise?

[sc94597]

Pemalite said:

sc94597 said:

I never said the X1 was customized for any specific form-factor. In fact, that is my original point. It was designed to be used in many different devices with many different purposes. Also, let's not become the pot calling the kettle black with "you have twisted my statements to try to fit your narrative." You have the tendency of splicing statements out of their original context and replying to them as if they aren't part of a stream of thought. You do this often and to many people. Besides, I didn't "twist your statement" I addressed that you put words into my mouth that I never made, like the idea that hardware was sitting somewhere and then was being used in Switch's.  

1. You stated Nintendo purchased the Tegra X1 as a "Hand me down". - It's a false statement.

sc94597 said:

Every single design decision requires a labor-force that typically makes six-figures per FTE (in the U.S) to plan, design, and test those decisions. Microarchitecture R&D is separate from the R&D for designing the actual chipsets. There were still FTE's that made decisions about what the specific SOCs look like with the X1, and they cost money to employ. This isn't some binary of "R&D is done when Maxwell [or other micro-architecture] comes out." Tons of decisions are made after that. 

2. Good thing the USA isn't the leader in chip manufacturing then.

sc94597 said:

And yes, planning the development of a chipset specifically with one platform in mind and its form-factor does require more R&D than just plopping in an already designed general-purpose chip, turning off a few CPU cores, and adjusting clock-rates. Are you seriously arguing otherwise here? 

3. The Tegra X1 is a standardized commidty part, like an SSD or Ram.

Nintendo purchases it, sets clocks and voltages as per their design specifications via firmware, then gets ANOTHER company to assemble it.

sc94597 said:

If I were to make an analogy, its the difference between making a new video game and making a remaster of a game. In both scenarios you might be using middleware like a game engine (in this analogy that would be the micro-architecture design that has already been done) but the latter (remaster : adjusting a few clock rates) requires a lot less work, FTE, and money than the prior (developing a new game : designing hardware to fit a specific form-factor and purpose over the course of years.) 

4. No one designs chips strictly for consoles anymore. They are assembly of already pre-designed parts.

The days of Cell are over.

sc94597 said:

The Switch 2 SoC isn't based on the T234 in the same way T210 in the Switch was on T210 in the Shield. I'm not sure where you got that idea. It was developed adjacent to it and the design decisions were made with its form-factor and intended use in mind. Some decisions might have been made for Ampere-based Tegra in general, but not all of these decisions. That's why it doesn't, for example, include specialized edge-compute hardware but has an FDE. Just because they both use the same basic architecture and share some design choices, doesn't mean there isn't different R&D being done for these two chips beyond the micro-architecture and common chipset family R&D. 

5. Tegra T234 was designed primarily for Nvidia's Jetson AGX Orin for Industrial/HPC applications.

Nintendo took the Switch approach and took an already designed nVidia chip for the Switch 2 and made some clockspeed and voltage changes to fit within it's design goals.

T234 being Ampere and ARM based is a quick time-to-market with low R&D investment as the architectures were already designed and ratified for other markets.

1. It is not false. The chipset in the original Switch SKU had the same exact design and was identical to the T210 in the Nvidia Shield TV.

It is not just a matter of using the Shield's chipset as a base and then making specialized modifications for a hybrid closed-ecosystem console. It was the same exact chip-design, but with a few CPU efficiency cores disabled (still present on the physical hardware though) and clock speeds under-clocked. These things, as you noted, would be done by Nintendo and not Nvidia. 

That is not how most console hardware, even today, inherits existing designs. The T239, for example, is a chipset designed and used exclusively by the Switch 2. There is no identical to x device in the Switch 2's circumstance. It inherits general design decisions from other Ampere and Tegra Orin chipsets, but it is its own chipset which had to be specifically designed by (an) Nvidia SOC R&D team(s), using constraints from Nintendo and with its intended use-case in mind. 

2. U.S companies are the leader in chip design (although not exclusively) and the design work does happen in the U.S. Why do you keep bringing up manufacturing and fabrication? Nvidia is not a chip manufacturer, they're a chip design company. Nintendo's contract with Nvidia is for chip design, and you know this because you say as much in your response here. The T239 would not exist if Nintendo didn't work with Nvidia to ask for it to be designed for them. Tegra Orin would just be a SOC family with one fewer SOC. The T210 would exist if Nintendo didn't contract with Nvidia for the Switch. That's the difference. That is what is meant by "hand me down." 

3. Yes, and the T239, comparatively -- is not. Which is my point. It's not for purchase by any other company than Nintendo. It was designed specifically for Nintendo and the Switch 2 with their input and constraints, working with Nvidia. No other device will use the T239. 

4. See points #1 and #3. There are no other devices that will use the T239. 

5. And Tegra T234 is not the chipset in the Switch 2. The Tegra T239 that is in the Switch 2, is a different SOC from the T234. It's not the same situation with the original Switch where the same exact chipset (T210) was popped into the Switch 2, the efficiency core cluster deactivated (not even removed) and frequencies changed. The T239 has specialized accelerators for gaming-specific tasks, different SM/core counts, and has the specialized edge-compute accelerators (NVDLA) removed compared to the T234. You don't see anything like this with the T210 vs. T210 (in another device) situation because that was an actual general-purposed chip meant to be used in tablets (amongst other devices) and not specifically designed for edge-compute as the T234. Nintendo was able to just purchase the T210 from Nvidia as-is, because it worked well enough. They couldn't do that with a T234 (for its size alone.) And even then Nintendo had to disable its efficiency core cluster that existed physically in the Switch, which they wouldn't have done if they had a chipset designed specifically for the Switch, as they do with the T239 for Switch 2. They'd just not include them from the start and have a different CPU configuration that is fully utilized by the Switch, if the Switch's chipset were semi-custom like the Switch 2's. This is why the T239 doesn't physically have the NVDLA; it's a semi-custom chip for the Switch 2, whereas the T210 was not for the Switch and it therefore had redundant/vestigial hardware (like the A53 cluster.) 

Very few Tegra chipsets in the last few Tegra families are general-purpose in the sense the X1 family was. Most are specifically designed for edge-computing, unlike the X1 family, which was genuinely meant to be used in a variety of devices, including consumer tablets and small form factor boxes. Nvidia basically abandoned the tablet/consumer market after X1, which is why Nintendo couldn't just pick a SOC design 'off the shelf' and say "good enough, we'll just disable some physically extant hardware" this time around. 

Last edited by sc94597 - on 04 May 2025