Maxwell or Pascal doesn't matter

Source

I know I'm pretty late and I know that most of you are probably sick of hearing Switch rumors, but the last piece of news about Nintendos upcoming console needs some addressing. Don't worry I'll make it quick.

So over the course of last week a rumor surfaced via Venture Beat that the Nintendo Switch won't use the actual generation of Nvidia chips dubbed ''Pascal'', but instead the older generation ''Maxwell'' Chips. If you want to read up on the rumor you can do this here at the Source.

Now, as expected, this rumor created a lot of buzz. Fears about repeating the mistake of the Wii-U surfaced. People were quick to jump to conclusions and call the Switch underpowered. And while I can't put your mind at ease, if you fear the same, I hope I can at least give you some insight into the topic.

The Maxwell/Pascal Difference:

First let me say that the difference between Maxwell and Pascal is almost entirely down to the manufacturing process. Maxwell is made on 28nm (TX1 even on 20nm), while Pascal is produced on 16nm.

The actual architectural difference between the two is minimal, and aside from an improved color buffer compression, completely irrelevant for the Nintendo Switch.

Still the article never talks about, or even mentions, the manufacturing process. For a believable leak this is pretty strange, because manufacturing is obviously the defining difference between the two set of GPUs.

Another problem of this article is that it gets the difference between the two chip sets completely wrong. Saying ''Nintendo's box is relatively small, and so it has to fit into the heat profile of a portable device, rather than a set-top box. That's another reason that explains the older Maxwell technology, as opposed to the Pascal's state-of-the-art tech.''

Pascal is literally a more power efficient version of Maxwell, so the incentive would be the other way around.

The auther then says ''we expect the Nintendo Switch to be more than 1 Teraflop in performance'', which is notably higher than even those people, who were expecting Pascal, were considering. If this is a Maxwell chip, then that would mean at least 4SMs (512 ''CUDA cores'') at 1GHz, because they wont be able to push much past that on 28/20nm. This would make the GPU much lager than anyone has expected.

What does this mean?

So does this mean we can all rest easy and the Switch will use the newest Pascal architecture? I'm afraid not. But in my mind there are a few possible scenarios.

The Nintendo Switch uses a custom made Maxwell Chip at 20nm, and simply has a much larger GPU than anticipated to account for the performance.

Nintendo looked at the feature set planned for the Pascal when design started, realized that the new features were largely irrelevant to their vision, and decided that they would save time and just use a Maxwell shrunk to 16nm instead. This would technically be a Maxwell Chip, but performance wise it would be completely identical to a Pascal GPU.

The sources are wrong about the 1 Teraflop performance, the Maxell, or both.

So in the end the only worthwhile thing to take away from this article is this quote:

''We expect the Nintendo Switch to be more than 1 Teraflop in performance.''

A Teraflop achieved by a Maxwell and a Teraflop achieved by a Pascal are identical. And to the customers it's irrelevant if this was achieved via using a lager Maxwell Chip on 28/20nm and at a lower clock or a smaller Pascal Chip on 16nm and at a higher clock.

In the end there was a lot of buzz for nothing, like always when a new Switch Rumour hits the Internet. Let's all hope that this changes at January, 12^th with the final presentation of the Nintendo Switch. And hopefully Nintendo won't be a secretive anymore in the future.

And as always

thanks for reading