I'd be more curious as to what the game engine sees since we can't expect every game to be patched to be aware.
20nm with the existing Tegra design would be a cheap way to test and see if it was successful and TSMC still has orders for various nodes.
Fairly difficult politics there, I think. One year after launch, you get a console twice as powerful. But only this revision. Other revisions won't get any faster, just cheaper. So the first purchases get lumbered with a naffy experience versus all the other buyers - In your face, early adopters! That's fine for hardware that gets frequent iterations and upgrades, where it's known, but it's lame for a console. That much of a performance advantage should be reserved for a mid-gen refresh a few years in, by and large.
The Switch doesn't have Dolby Digital / DTS encoding for surround?
20nm and 16nmFF is similar in density and 20nm is cheaper per finished wafer. Can’t really see 16nm being a very desireable option from Nintendos perspective. 7nm gives, at high volume, the option of both drastically reducing die area while simultaneously improving performance and performance/W. That process is entering volume production pretty much now, and since the Switch is selling well, it may be a decent option to actually produce a custom SoC. If they do, performance predictions are difficult.
Glass touchscreen would be required for me to even consider it. Shitty plastic getting all scratched up...no way.
I assume you are talking quite a long time in future though, at least 2 years.
You are quite correct. It’s not that there wouldn’t be technical benefits to moving to 16nmFF, it’s that it doesn’t seem to make much economic sense. It would simply cost more per chip, and even if Nintendo could save a bit on the construction due to lower power draw, that’s not enough to help them drop prices. I don’t think adding a bit of performance (and adding SKUs and associated issues) is a priority for them. Once there is a lithographic node that lets them tick more boxes, or substantially change the capabilities of the device. A mid-life kicker is a possibility of course, but I’d estimate that at least three years after launch, or 2020. At that point TSMC is likely to be very comfortable with 7nm, and Apple will be shifting their new chips to 5nm freeing up fab capacity.
Although these things aren’t trivial, 20SoC probably wouldn’t be picked for a new design, something like 16FFC would probably be picked instead. However, for an existing design, staying on a (cheap) process can make a lot of sense. It saves on design expense, respins, wafer cost. Again, it’s not that a new process wouldn’t bring benefits technically, the question is whether a change in process can bring sufficient benefits in the marketplace to more than offset the costs involved, improving profitability over staying put.
TX2 is larger and made on a more expensive process.
