By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Forums - Nintendo - How Will be Switch 2 Performance Wise?

 

Switch 2 is out! How you classify?

Terribly outdated! 3 5.26%
 
Outdated 1 1.75%
 
Slightly outdated 14 24.56%
 
On point 31 54.39%
 
High tech! 7 12.28%
 
A mixed bag 1 1.75%
 
Total:57

We still don't know if the SOC is 5 or 8nn process? It seems too powerful for 8nm but did N really spend money on Samsung 5nm?



Around the Network
numberwang said:

We still don't know if the SOC is 5 or 8nn process? It seems too powerful for 8nm but did N really spend money on Samsung 5nm?

Could be dock mode is using a lot more power than we think. If they loosened the TGP to 20-25W and got some efficiencies with node maturation they could pull off the semi-confirmed frequencies. 

If they knew they were going to release the platform so late, they probably would've gone with a Lovelace chip and be able to have something more powerful than the Series S in many ways when docked (assuming the TGP is loosened.) 

I still think it is probably 5nm, but the low battery life makes me wonder. 



EricHiggin said:

SW2 hardware capabilities and game graphics vs PS5 hardware capabilities and game graphics, costs compared, Nin is expensive vs SNY. That's not a good comparison for Nin.

I'd say they're pretty evenly matched for the value:  The PS5 is expensive because of powerful tech, Switch 2 is expensive because it's highly capable for its form factor.



sc94597 said:
Biggerboat1 said:

You seem to be talking about docked. When sighting S2 being slightly weaker than Steam Deck I was talking about undocked, which if true would be a bit disappointing as the Deck is over 3 years old now... It's true that it launched for a bit more than S2 (at 256GB) though Nintendo will have vastly more bargaining power so should get a lot more bang for their buck due to volume (think the Deck is around 4 million sold, so under 1.5m per year).

Thankfully due to S2 assumed large install base, games will be optimized more than on the Deck, but it would have been nice to have both - Optimization + undocked performance that handily beat a 3yo Deck. Also, for CP specifically, this has been Nvidias showpiece for various technologies so it should run better than vs AMD hardware at a similar performance.

Maybe you have an insight into how much more a node step up would have cost? They'd have saved money from not having to change node mid-gen (if they followed S1 strategy) and they'd also presumably have gotten another year or 2 out of the hardware as it would have delayed the point at which the visuals would be seen as obsolete.

I'm just going from what DF is saying, so if that proves inaccurate then I'll obvs change my position.

In terms of raw raster the Switch 2 is about 55% as powerful in handheld mode as in docked mode. That roughly aligns with the difference between the Steam Deck's and Switch 2's targeted modes.

The Switch 2 is targeting 2 times the output resolution at the same frame rate and settings docked as the Steam Deck. I am not going to talk about internal resolution because I don't think the Switch 2 is going to get to 1080p without DLSS in the game and we don't know what sort of DLSS solution they'll use. 

So there is no reason that if they hit the 1080p 30fps target docked they can't hit a 720p 40fps or slightly higher resolution 30fps target in handheld with similar settings to the Steam Deck/PS4/Switch 2 (Docked.) 

When it comes to projections (rather than analysis of existing products) I don't find Digital Foundry to be very reliable. Remember, they originally thought the docked Switch 2 would be less performant than the Steam Deck. That is obviously far from the case. They're questioning a highly capable team's ability to achieve perfectly doable frame-rate and resolution targets given what we know of the system. If CDPR thinks it is feasible to have a 1080p 30fps mode and a 40fps at some lower resolution, then I think it is more likely than not to happen. And if that is the case, then the handheld mode will almost certainly be more or less match to match to what the Steam Deck outputs. 

I guess time will tell. You're obvs much more knowledgeable about the underlying tech.

I don't think DF have an axe to grind though, they seem to be calling it how they see it.

They also seem pretty sure it's 8nm, going by the reported dimensions of the board.

CDPR are obviously very talented but they made a bit of a dog's dinner of the PS4 & Xbox One port so who knows, again let's hope for the best & see.

You mentioned in your post to numberwang that Nintendo could have gone with Lovelace over Ampere.

If DF are correct and we've ended up with an 8nm Ampere when we could have got a 5nm Lovelace setup you can maybe understand my slight deflation as that would have been a serious step up.

I can't help but think that if it was MS or Sony's project they would have made that level of hardware happen whereas Nintendo always seem to go a bit mid...

A big part of Nintendo's audience have grown up & are likely more tech savy than generations past & I think more ambitious hardware would have gone down well & increased the longevity of the system. I know I've tailed off on using my OG Switch as I preferred to wait & play games like TOTK & even Link's awakening on hardware that could comfortably handle it.

I'm not sure how much they'd have had to charge but personally I'd have been happier to pony up an extra 25 to 50 bucks if it meant cutting edge vs what we've seemingly ended up with, which just seems 'fine'.



EricHiggin said:
curl-6 said:

If you reduce the price of your product you have to either reduce the manufacturing cost, (cheaper hardware) or eat a loss for every sale.

If reports are to be believed, they're already going to be taking a loss on hardware, so losing a further $50 per sale isn't something any company would do unless they could make up the difference somehow.

SNY charged less for PS5 and didn't make it weaker.

Losing money upfront on hardware is the norm, and only recently has SNY and MS tried to break even or make a small profit at launch. Nin is a bit different as they try harder to make money right off the bat with hardware, but that's not easy to do if you're offering more state of the art tech. 

Make up the difference somehow, like sell games, lots of games, expensive games?

Sony didn't "charge less" for PS5; they charged what they thought would make them the most money; they took a loss on hardware (initially) because they make it back by having more people to milk for subscription money.

Nintendo isn't as focused on subscriptions and the like, and they don't have a ton of other huge non-gaming divisions to draw money from like Sony and MS do.

Switch 2 is already apparently sold at a loss, no company on earth would eat a further loss if it didn't benefit them.



Around the Network
EricHiggin said:

SW2 hardware capabilities and game graphics vs PS5 hardware capabilities and game graphics, costs compared, Nin is expensive vs SNY. That's not a good comparison for Nin.

Switch 1 launched at the same price as the PS4 and was less powerful, yet people clearly saw the Switch as better value since it outsold the PS4 by miles.



Biggerboat1 said:

If DF are correct and we've ended up with an 8nm Ampere when we could have got a 5nm Lovelace setup you can maybe understand my slight deflation as that would have been a serious step up.

Fun fact - 4050 mobile vs 3050 mobile (2560 version, so both with same 2560:80:32:80:20 config) is some 30% faster on average for same power draw.

That's what 4nm for 4050 vs 8nm for 3050 does, basically allowing for much higher clocks in 4050. So yeah, it would be really nice if Nintendo went with smaller node, which is not very likely.



HoloDust said:
Biggerboat1 said:

If DF are correct and we've ended up with an 8nm Ampere when we could have got a 5nm Lovelace setup you can maybe understand my slight deflation as that would have been a serious step up.

Fun fact - 4050 mobile vs 3050 mobile (2560 version, so both with same 2560:80:32:80:20 config) is some 30% faster on average for same power draw.

That's what 4nm for 4050 vs 8nm for 3050 does, basically allowing for much higher clocks in 4050. So yeah, it would be really nice if Nintendo went with smaller node, which is not very likely.

Exactly. If the OG Switch had been 30% stronger then perhaps I and others wouldn't have held off on purchasing TOTK & other software and Nintendo would have had more money on the back-end to offset the higher initial investment (not saying that they had an option to achieve this with OG Switch, just pointing out S2 could suffer a similar fate).

My fear is that DLSS may be too expensive in a lot of situations for S2 (this was touched on by DF in their Cyberpunk analysis), which would be a major bummer...

30% more grunt would have given them the extra headroom and then some.



archbrix said:
EricHiggin said:

SW2 hardware capabilities and game graphics vs PS5 hardware capabilities and game graphics, costs compared, Nin is expensive vs SNY. That's not a good comparison for Nin.

I'd say they're pretty evenly matched for the value:  The PS5 is expensive because of powerful tech, Switch 2 is expensive because it's highly capable for its form factor.

curl-6 said:
EricHiggin said:

SW2 hardware capabilities and game graphics vs PS5 hardware capabilities and game graphics, costs compared, Nin is expensive vs SNY. That's not a good comparison for Nin.

Switch 1 launched at the same price as the PS4 and was less powerful, yet people clearly saw the Switch as better value since it outsold the PS4 by miles.

Based on what I was replying to before, I meant it as the hardware directly being able to push the software, and the software needing beefier hardware. I wasn't taking portability into account with that statement. However if you were taking overall functionality into account, then yes, it evens the playing field in terms of overall usefulness, assuming the mobility is important if not key to those using the device. For those who leave it docked, then that wouldn't be as true for them.



PS1   - ! - We must build a console that can alert our enemies.

PS2  - @- We must build a console that offers online living room gaming.

PS3   - #- We must build a console that’s powerful, social, costs and does everything.

PS4   - $- We must build a console that’s affordable, charges for services, and pumps out exclusives.

PRO  -%-We must build a console that's VR ready, checkerboard upscales, and sells but a fraction of the money printer.

PS5   - ^ -We must build a console that’s a generational cross product, with RT lighting, and price hiking.

PRO  -&- We must build a console that Super Res upscales and continues the cost increases.

curl-6 said:
EricHiggin said:

SNY charged less for PS5 and didn't make it weaker.

Losing money upfront on hardware is the norm, and only recently has SNY and MS tried to break even or make a small profit at launch. Nin is a bit different as they try harder to make money right off the bat with hardware, but that's not easy to do if you're offering more state of the art tech. 

Make up the difference somehow, like sell games, lots of games, expensive games?

Sony didn't "charge less" for PS5; they charged what they thought would make them the most money; they took a loss on hardware (initially) because they make it back by having more people to milk for subscription money.

Nintendo isn't as focused on subscriptions and the like, and they don't have a ton of other huge non-gaming divisions to draw money from like Sony and MS do.

Switch 2 is already apparently sold at a loss, no company on earth would eat a further loss if it didn't benefit them.

Again, SNY wanted to charge more for PS5, but decided against it. They also had multiple designs prepared. They didn't choose the weaker design for cheaper, or even if the PS5 we got was the cheapest design, then they should've delayed the launch way early, and downgraded the hardware based on your thought process, but they didn't.

Nin is moving more and more towards subscriptions. You don't make a chat button and then lock it behind a sub (after a year) if you're not looking to grow and profit from your subscription business.

MS constantly ate it with XB in the earlier gens, and then early on with GP. SNY ate PS3 early on, and they ate it hard. Did those losses "benefit" them? Even in the long run? That's not to say Nin should put themselves in that situation right now, obviously not, but Nin as a company, and SW2 as a device, are nowhere near that realm of hurt, which means there's wiggle room.



PS1   - ! - We must build a console that can alert our enemies.

PS2  - @- We must build a console that offers online living room gaming.

PS3   - #- We must build a console that’s powerful, social, costs and does everything.

PS4   - $- We must build a console that’s affordable, charges for services, and pumps out exclusives.

PRO  -%-We must build a console that's VR ready, checkerboard upscales, and sells but a fraction of the money printer.

PS5   - ^ -We must build a console that’s a generational cross product, with RT lighting, and price hiking.

PRO  -&- We must build a console that Super Res upscales and continues the cost increases.