Nintendo Switch Technical discussion [SOC = Tegra X1]

[Goodtwin]

What mattered for the port discussion were the scarcity of AAA 3rd party franchises that move tens of millions across a generation, and those were effectively scarce.
The huge gap in performance resulted in the Switch getting no Call of Duty, no Battlefield, Battlefront, no recent Final Fantasy, no Overwatch, no Far Cry, no recent Assassin's Creed, no Destiny, no The Division, no GTA, no Red Dead Redemption, etc.
The list goes on and on.

Ok, I think I am on the same page as you now. Most AAA developers looked at the potential return on investment, and decided that it wasn't worth the investment. This was my point of view from the beginning, it would and has been a financial decision, and not an insurmountable challenge that couldn't be overcome. Some publishers like Bethesda have looked at Switch as an opportunity to grow their audience. Probably willing to accept smaller returns on investment inorder to gain new fans. Then you have Capcom who has supported Switch, but with very low investment, primarily selling old games. This strategy had worked pretty well for them because the portability factor is enticing for many.

Good comments on FP16. I miss the good old days when Shin'en would give in depth interviews detailing things like that.

Nintendo is likely to stick with this hybrid setup for its next console, meaning Nintendo will again be under powered in a big way. Even releasing 4 years after the PS5 probably wouldn't be long enough for mobile tech to match its capabilities in the portable form factor. For the successor, I think the smart move would be to release once Nivida can create a new Tegra chip that exceeds the base PS4. As we have seen with Switch, ports of old 360/PS3 games have done pretty well on Switch, and I do not see why PS4 ports on Switch's successor wouldn't be well received in the same fashion. Question is how long will it take Nvidia to get there.

 

[swaaye]

I think the ports are happening because Switch is portable and people are willing to buy their favs again in a portable format. If it was not portable, I think Switch would be bombing like the Wuu did.

It's also new for a portable to be able to run games of such sophistication. I think that's an exciting prospect / gimmick for people.

Selling games again in a portable format is historically commonplace.

 

[Deleted member 13524]

Ok, I think I am on the same page as you now. Most AAA developers looked at the potential return on investment, and decided that it wasn't worth the investment. This was my point of view from the beginning, it would and has been a financial decision, and not an insurmountable challenge that couldn't be overcome.

Erm.. No, I wrote the opposite:

Devs need to take away too much of their 9th gen games to make it work on the Switch.

It's not only about effort, it's about devs not willing to deface their games to run on the Switch.
It's not because it's too much work, it's because the console simply can't run them at their intended quality.​

I think the ports are happening because Switch is portable and people are willing to buy their favs again in a portable format. If it was not portable, I think Switch would be bombing like the Wuu did.

It's also new for a portable to be able to run games of such sophistication. I think that's an exciting prospect / gimmick for people.

Selling games again in a portable format is historically commonplace.

Nah.
The Vita got Borderlands, DoA5 and NFS Most Wanted.
Running "grown up" games on a mobile platform isn't something Nintendo invented.

 

[swaaye]

Nah.
The Vita got Borderlands, DoA5 and NFS Most Wanted.
Running "grown up" games on a mobile platform isn't something Nintendo invented.

Something is different with Switch though. The original PSP actually had a fairly complete port of sorts of Oblivion coming but it was canned. Now we get full portable Skyrim with Switch and loads of other big games too.

Sony portables just didn't capture the mass market appeal or interest of publishers for some reason. Something was wrong with the marketing there.

 

[Deleted member 13524]

Something is different with Switch though.

Of course it's different. Sony quickly redirected all of Vita's marketing and 1st-party development efforts towards the PS3 because smartphone/tablet gaming was on the rise and they thought they wouldn't stand a chance.
The Switch is now Nintendo's plan A and plan B. There's no DS/Gameboy in case the home console fails, or vice versa. A failing Switch means Nintendo needing to scramble and launch another console fast or turning into a software developer.

The original PSP actually had a fairly complete port of sorts of Oblivion coming but it was canned.

It seems obvious to me why it was canned. There's just no way a game like oblivion could have a decent gameplay with only one analog stick on the left side. The NDS used the touchscreen to replace the right analog on its first person games like Metroid.

 

[Shifty Geezer]

There's also the case of storage. Loading for PSP was from UMD. The 1.8 GBs would have been enough to squeeze in the cut down assets from the 5 GBs original, but streaming from that optical drive versus the HDDs of the console and PC was of course never going to happen. The slow speed of UMD was going to limit the types of games that could be ported.

The important point here is, as mentioned many times before, any game can be ported to any system. It's just a matter of what degrees of compromise you need to make, and a subjective consideration as to whether the end result is still the sae game or not. The presence of ports of games on NSW don't indicate anything about its technical prowess, but its economic viability to take ports and turn a profit on the work needed in making a game fit. The technical prowess of the machines is in considering how those ports fare versus other hardware.

tl;dr : Can we please stop talking about NSW getting ports in this thread and instead just talk about the technical features of those ports.

 

[Entropy]

tl;dr : Can we please stop talking about NSW getting ports in this thread and instead just talk about the technical features of those ports.

No, we need to dig up the posts where authorative statements about how those ports would be impossible on the Switch and publicly shame those individuals into humility.
As if.
Anyway, Goodtwin raised the question of when Nvidia could supply Nintendo with a chip that would be roughly on par with the PS4, and I’d estimate that would likely be around 2021 once the 5nm node has been used for volume production for a year or so, and LPDDR5 has had a couple of years of volume production. But that something is technically feasible doesn’t mean that it will be done. Nintendo is in a better position to act with greater confidence in terms of procurement than they were when moving forward from the Wii-U which created a lot of doubt everywhere about their future viability. But what do they want to do? And when? I could make a case for such a device at that time, because it would give them easy access to the entire catalogs of XB1/PS4 and any cross generational titles, and they would be able to run anything produced for the next generation at todays mainstream stationary resolutions. In fact, I could see portability, petite dimensions and absence of noise/power draw concerns winning out over the graphical refinement of the new stationary consoles and Nintendo dominating the overall console market. While for most on these forums the Switch is a second complementary platform, for many it already serves as a one stop solution for mobile and stationary gaming ”needs”, and a really strong well timed entry could change the current landscape a lot.
But Nintendo could have other plans, or prolong the lifetime of the current environment to 3nm or... They have a lot of options.
We just don’t know.
But if they are smart they will be talking to the big publishers about their future plans - and the big publishers would be fools not to talk to Nintendo.

 

[Deleted member 13524]

But if they are smart they will be talking to the big publishers about their future plans - and the big publishers would be fools not to talk to Nintendo.

Why?
Has Nintendo made big publishers a lot of profit within the past 15 years? Is the Switch bound to change that anytime soon?

Big publishers will jump on Nintendo's next console if it can play the games they have in the pipeline for the other two consoles.
Until Nintendo successfully shakes off the general sentiment that their consoles only serve to sell their games, 95% of what they're getting are the cheap ports of sports titles that use the N-1 gen or N-2 gen engines with updated rosters.

 

[Goodtwin]

Cartridge pricing would be interesting to know. Most publishers/developers have stayed away from anything higher than 16GB. Dragon Quest Heroes used a 32GB cart pretty early on in Switch's life cycle, but the Witcher 3 will be the only other game to use it as far as I know. I cant imagine they cost more than $6-8 at this point.

I would like to see Nintendo release future Switch models with more on board memory. It seems to me that having more space would entice more consumers to consider buying more digital content, and digital content brings higher profit margins. You can buy 256GB Micro SD cards sell retail for $40 now, I would have to think that NAND memory would be cheaper than Micro SD Cards, especially in the volume that Nintendo purchases.

Selling games again in a portable format is historically commonplace.

Very true, and may ultimately be the primary business model for third party publishers with Nintendo's hybrid platform. Switch is able to outperform last gen consoles by a decent margin, so not only do these ports offer portability, but also a modest increase in visual quality compared to previous console iterations.

Why?
Has Nintendo made big publishers a lot of profit within the past 15 years? Is the Switch bound to change that anytime soon?

Yes, Capcom made boat loads of money with Monster Hunter on 3DS. Activision, EA, and Ubisoft all did well on the Wii for the first few years of its life. A ton of smaller third party publishers did very well on the 3DS. The Switch will certainly be a small piece of the pie for Activision, Ubisoft and EA. At the very least, Nintendo Switch is a platform to release last generation games, where porting cost are minimal, and even modest sales results in profits in the millions. You also have small budget games like Project Octopath from Square Enix that have done very well, and their RPG Factory teams are full steam ahead with multiple projects in development. Then we have Indies on the eshop, many of whom have had far greater success on the Switch than the PS4 and X1.

 

[swaaye]

It seems obvious to me why it was canned. There's just no way a game like oblivion could have a decent gameplay with only one analog stick on the left side. The NDS used the touchscreen to replace the right analog on its first person games like Metroid.

Yeah a single analog stick would be annoying but I'm not sure it's a show stopper. Input scheme horrors aren't all that uncommon with console gaming.

But what was Sony thinking having only one analog stick at that point in time? Yikes. Its scheme is worse than an N64.

 

[Entropy]

Why?
Has Nintendo made big publishers a lot of profit within the past 15 years? Is the Switch bound to change that anytime soon?

If the major publishers don't publish on Nintendo platforms they won't get revenue from them either. Which isn't a problem if the platforms they support have vastly greater volume. But the reality of the situation is that the Switch sells well (the best, actually, in both Japan and the US, I don't know Europes numbers), and software on the platform sells really well. Do you think they want in on that or would they prefer to let other publishers take that revenue uncontested?

As for the future, let me exemplify. I'm currently winding down work heading into a holiday, weather is nice, and yesterday for once I had the opportunity in the evening to sit on the porch, drink something refreshing, smoke a cigar, and just relax. As it happened, I had to do some shopping and noticed Final Fantasy XII for sale for the Switch, and picked it up just to check out what FF was about. (Riding on birds always seemed ridiculous to me, and I've never liked the Japanese penchant for mixing in the ridiculous and cute into otherwise rather dark adventures. Nor their grinding.) So I sat on the porch with my refreshing drink, my cigar and my Switch, playing a game from Square Enix that I haven't been interested in buying for years.
Note that I have a gaming PC that already performs at the level of the upcoming next generation consoles. But I couldn't bring that with me to where I actually wanted to be, nor could I bring a PS5 or XBoxWhatever. And if in the future the graphical compromise vs. stationary TV-attached consoles grows more esoteric, the stronger relatively the ergonomic benefits of the Switch format.

I'm not making any claims as to what Nintendo will actually do. They march to the beat of a different drummer, and to a large extent I think they are more concerned about how to thrive in a gaming environment that is ever more dominated by the mobile platforms than worried about whatever path Sony and MS pursue. From a technical point of view they have several options.

 

[Silent_Buddha]

Why?
Has Nintendo made big publishers a lot of profit within the past 15 years? Is the Switch bound to change that anytime soon?

Big publishers will jump on Nintendo's next console if it can play the games they have in the pipeline for the other two consoles.
Until Nintendo successfully shakes off the general sentiment that their consoles only serve to sell their games, 95% of what they're getting are the cheap ports of sports titles that use the N-1 gen or N-2 gen engines with updated rosters.

I'd consider UBIsoft and Bethesda fairly large publishers. And they have made quite a lot of money on the NSW. Mario + Rabbids (UBIsoft) is the #13 best selling title for the platform doing better than some of the Nintendo developed exclusives. Minecraft is 17. Skyrim 22nd. Heck, even Indie games are charting up where.

And now that Nintendo are more willing to allow 3rd parties to use Nintendo IP characters (An indie game just came out using Nintendo IP that's pretty good), it's even easier for 3rd party developers to make money on the Switch. Assuming the developer/publisher wants to work with Nintendo.

The NSW is light years ahead of the Wii and Wii-U WRT to developing games on the platform, especially if the goal is games that are similar to what is on PlayStation and Xbox. It doesn't require nearly as much effort to port to the NSW as it was to try to shoe horn games onto the Wii or Wii-U. That doesn't mean it doesn't require a lot of effort, just that it's significantly easier than those past platforms.

Regards,
SB

 

[RDGoodla]

I can’t remember Nintendo has ever invested in shrinking SOC with new process.

So I really doubt how nintendo designs a pro version of SWITCH. And what will be their next-gen SOC if Nvidia has no new Tegra?

 

[AlNom]

Wii is effectively a 90nm shrink of GameCube with a +50% buff to alacrity.
WiiU is some Legendary Unique pick-up that includes previous buffs but you find it when your character is 5 levels ahead already.
Switch is one of those baka items you find in a shop.

pica pica200
pica choose XL



----

They ought to be okay if the actual main market of Tegra suddenly dries up and nVidia doesn't do all the R&D for automotives beforehand.

 

[Shifty Geezer]

And what will be their next-gen SOC if Nvidia has no new Tegra?

There's already a newer Tegra, X2. That'd serve perfectly as a Switch Pro.

 

[Goodtwin]

There's already a newer Tegra, X2. That'd serve perfectly as a Switch Pro.

Isnt the Tegra X2 much larger than the Tegra X1? If there were to be a Switch Pro model, I would think the easiest solution would be to move the Tegra X1 to a smaller process, and increase clock speeds. I would assume there is a faster LPDDR4 memory chip available by now? Nvidia will probably fully customize a chip for the Switch successor, but doubt that is in the cards for a mid gen refresh. I am skeptical on a Pro model being released. The mini model is pretty much a sure thing, especially with the recent leak of Switch Mini accessories being manufactured in China.

 

[Deleted member 13524]

There's already a newer Tegra, X2. That'd serve perfectly as a Switch Pro.

There's an upgrade path for a New Switch with Tegra X2, but there's no upgrade path for an actual successor.
Maybe Nintendo will be willing to go with Qualcomm ARM+Adreno or Samsung ARM+Radeon for the Switch 2, but with nvidia providing all the dev tools for the Switch, they might be rather stuck here.

At least Nintendo is less anal about backwards compatibility now than they were up until the Wii U, so they may not feel as pressured to keep using a Geforce GPU.

Isnt the Tegra X2 much larger than the Tegra X1?

Looking at nvidia's gigantic Turing chips, 16nm seems to be rather cheap right now. A Tegra X2 right now may not be more expensive to make than a Tegra X1 was back in late 2016.

Biggest problem would be RAM. TX2 uses 128bit LPDDR4, twice as wide as TX1 requiring twice as many x32 LPDDR4 chips. Going with TX1's 64bit would probably hinder the performance advantage of TX2, but if all they want is to provide docked performance @ mobile mode then they can probably just stick to
64bit.

 

[AlNom]

What's the state of ARM CPUs nowadays? Anything with a sort of ... idk, binary compatibility for A57 edit: A72 ->A75? (Seemed like A72 was exactly what a die shrunk A57 would be, while A75 bolts on some L3)

I wonder if Nintendo shouldn't just custom order something along with Turing, and go cache-heavy anyway. Switch is a proven seller now (vs coming off of the failed Great Experiment that was WiiU)

/expecting just "12nmFF" TegraX1 :V

 

[swaaye]

Considering the Switch is pretty popular maybe a fully custom chip makes sense now. Using Tegra X1 was a low risk option at the time. Those newer Tegras are loaded with useless functionality for a game machine.

 

[Goodtwin]

Considering the Switch is pretty popular maybe a fully custom chip makes sense now. Using Tegra X1 was a low risk option at the time. Those newer Tegras are loaded with useless functionality for a game machine.

That's why I think Nintendo will be willing to invest money for a customized chip with the Switch successor. Tegra X1 was pretty much a steal for Nintendo. Didn't have to spend big bucks for a custom chip, and it was still one of the best performing mobile processors on the market when Switch launched. Even if the roadmap ends for the Tegra line of processors, Nvidia still shifted to a power efficiency focus for its GPU's. So really, matching up some ARM CPU cores with whatever their current GPU cores are in a few years shouldn't be too tough for Nvidia. If they do want to maintain backwards compatibility they could always include the 4 A57 cores on the chip.

https://www.anandtech.com/show/13084/samsung-announces-first-lpddr5-64gbps-data-rates

If Nintendo does want to create a Pro Switch model using a shrunk TX1 at higher clock speeds, they could pair it with the newer LPDDR5 memory, offering plenty of extra bandwidth to accommodate the higher clock speeds.