kevboard
Member
so good couldnt see dlss and said switch 2 to week to run it ahahahahaha
![]()
you're still on your DF Derangement Syndrom trip I see.
so good couldnt see dlss and said switch 2 to week to run it ahahahahaha
![]()
Some reality check for sure. 1500 CUDA cores vs 2500 in 3050 so don't put a lot of faith in DLSS beyond FHD.
That being said it is floating more or less around Steam Deck level of performance.
never broughted him up buuuut he was wrongyou're still on your DF Derangement Syndrom trip I see.
so good couldnt see dlss and said switch 2 to week to run it ahahahahaha
![]()
The efficiency and features set are not the same… There's a improvement on every new gen.literally every single Nvidia RTX GPU is capable of every single DLSS version.
you can run DLSS4 on an RTX2060 if you want to.
this actually doesn't change anything...
Actually, no. RTX 20 series uses Touring architecture that's two gen older than Lovelace available the RTX 40 series.So, Digital Foundrys analysis, insisting it was a 8nm part and performance estimates, using a heavily downclocked RTX 2050 GPU, was pretty much spot on...well done, Mr Richard Leadbetter!
Actually, no. RTX 20 series uses Touring architecture that's two gen older than Lovelace available the RTX 40 series.
This is everything you need to know. Switch is capable of DLSS 3.0… The secret sauce of the Switch 2.
![]()
Well, this is something I don't get... Nvidia just decided to ditch OFA for FG after Switch 2 was designed and there's also FSR FG which is equally good so I don't think Ada features are still required for thatThe leaks of the Matrix demo mentioned DLSS 3.1
Chip printed date says its design is from 2021 tho, are we sure Ada was ready enough by then to secure it being so dominant over Ampere in T239?Lovelace is no slightly better is Way better… The CPU is way better too, the issue with the 8-10nm is going to be the power consumption in handheld mode. Other than that is this leaks are true, i better than the leak that we have before. If the GPU is Lovelace architecture over Ampere is a solid Win…
2050 isn't Touring, it's Ampere.Actually, no. RTX 20 series uses Touring architecture that's two gen older than Lovelace available the RTX 40 series.
Well yes the 2050 indeed is Ampere… still a Gen older.2050 is Ampere based.
The efficiency and features set are not the same… There's a improvement on every new gen.
Remember that Nvidia confirmed is a custom made GPU specifically made for the Switch 2.Chip printed date says its design is from 2021 tho, are we sure Ada was ready enough by then to secure it being so dominant over Ampere in T239?
So it's exactly as was leaked 5 years ago lol
Yes, and the printed year says 2021, so unless that means nothing at all, I'd say Ampere is the base and Ada features are complementing it.Remember that Nvidia confirmed is a custom made GPU specifically made for the Switch 2.
It's not about architecture it's about chip layoutThis is the old leak. That GPU was Ampere this leak is different
![]()
Only Nvidia knows that because is a CUSTOM MADE GPU, so, there's nothing comparable in the market. So, using GPUs available in the market to compare is a waste of time.Yes, and the printed year says 2021, so unless that means nothing at all, I'd say Ampere is the base and Ada features are complementing it.
Edit: And BTW if Ada was the base why would it have anything from Ampere? It makes no sense, except the above being the actual case.
which still was being worked on and customized by them so its partially accurate. its the first hint we had and definitely pointed in the correct spotThis is the old leak. That GPU was Ampere this leak is different
![]()
It's not about architecture it's about chip layout
That leaks is from 2021, So, is an early prototype. Lovelace is from 2022.which still was being worked on and customized by them so its partially accurate. its the first hint we had and definitely pointed in the correct spot.
You missed out the line above.This is the old leak. That GPU was Ampere this leak is different
![]()
No, but Nvidia himself said that's a custom design GPU. So, having Lovelace feature set is great news and probably explains the great results that we are seeing in some games.You missed out the line above.
Watch the videoLol…
No, but Nvidia himself said that's a custom design GPUs. So, having Lovelace feature set is great news and probably explains the great results that we are seeing in some games.
Also architecture improvements and optimizations.![]()
By feature set, do you mean frame gen?
Remember 5080 is on 4090 Level.Remember that Nvidia confirmed is a custom made GPU specifically made for the Switch 2.
The architecture is still Ampere though. Or are you ignoring the video for some reason?Also architecture improvements and optimizations.
Read again…Worse than Steam deck in handheld mode bruh
This… Lovelace is the news here.Nothing very new. Most of those specs were known for months/years.
A bit more than a PS4 in portable mode with a better CPU, 2/3 of a series S in docked mode with a worse CPU but more RAM...
But with a very modern feature set, allowing it to punch way above its weight. Good stuff overall, especially considering the diminishing returns in visuals this generation. I hope that they'll keep on upgrading dlss on this machine, and I wouldn't be surprised if traded blows with the Series S once developers get more familiar with it.
The architecture is still Ampere though. Or are you ignoring the video for some reason?
Current gen aren't PS4.Even is Ampere based with Lovelace features set (Custom made) is way better than current gen console GPUs … The Switch 2 can punch above his weight. And that was i expected since the beginning, PS4 handheld and PS4 Pro/Series S docked.
I am talking about the GPU… Not performance.Current gen aren't PS4.
lots of room there for future revisionsVery likely SEC8N based on gate pitch and die size
58%2/3 of a series S in docked mode
You aren't making any sense. AMD APUs in the current gen consoles support frame gen, that's the feature that 40XX has over 30XX (and this is a custom Ampere GPU).I am talking about the GPU… Not performance.
AMD frame gen is software based frame interpolation, Nvidia is hardware based… Something like FSR vs DLSS. That's means that Nvidia uses the tensor cores for that job not the CPU.You aren't making any sense. AMD APUs in the current gen consoles support frame gen, that's the feature that 40XX has over 30XX (and this is a custom Ampere GPU).
we will see performence comparison bookmarked oohoohoohooWorse than Steam deck in handheld mode bruh
The moving goalpost is not going to endwe will see performence comparison bookmarked oohoohoohoo
Going by Star Wars outlaws it's a little stronger than the deck in handheld, and a lot stronger when dockedSome reality check for sure. 1500 CUDA cores vs 2500 in 3050 so don't put a lot of faith in DLSS beyond FHD.
That being said it is floating more or less around Steam Deck level of performance.