There are several tests of Control from when DLSS 2 launched a few years ago, which get posted here a lot. Maybe others can link something more recent or with a higher output res?
Most console games do not do CPU and GPU work in the same shared frame time. They do CPU work for the current frame while GPU work is being done for the previous frame.
More important, I think, is not to make hurtful jokes. Or any jokes at the expense of other users.
DLSS can run in parallel with the next frame being rendered, adding at most a frame of latency.
This!
They typically don't reveal things at investor briefings, though they do rarely announce Directs during them.
Yes, the original Switch was revealed just under a week from their October 26th Earnings Report. In 2020, that was our main way of knowing when some things would be announced, like Pikmin 3 DX, and the July Partner Showcase. They also had an Indie World just a day or so before/after their Earnings Release last year.
Close to, yes. During is less common, but that is how they announced NX and its release window.
Oh dear….
To be fair, I think most of us aren't expecting it to be announced DURING the investor meeting. Close to it yes, and yes we have past examples.
Wasn't it specified to be 4k60 with dlss?
Sure, that’s true. Based on the various similar papers we’ve seen, the deepest hidden layer in DLSS likely only has around 256-1024 filters. It seems unlikely to me that there would be sufficient redundancy among those filters that you could remove, say, 10-20% of them just for the Switch without a significant impact to the output quality, even with retraining. But I’ll admit, it’s possible.
Not specified, although some have suggested DLSS is likely to have been used.
The 3DS was first announced a week before the end of the 2010 fiscal year via press release, the Wii U was first announced alongside Nintendo's April 2011 financial report, and the Switch was first announced as the NX alongside Nintendo's the announcement of their DeNA partnership in March 2015. While they used to be more commonly announced at events like E3, as Nintendo has since moved further away from live expos and more to their own Direct showcases, all three of their most recent major hardware releases have been first brought up around significant times in Nintendo's fiscal year.
www.eurogamer.net
Excuse me but, what were your expectations at the first place? This is a close approximation to the chip in the console lacking memory, console optimization and using Windows... DF himself stated this is all on the lower end of the spectrum, since T239 is a customized chip for a reason, the very fact it's already performing this incredibly well does all but lower expectations. And keep in mind this is under handheld clocks, docked should perform very similar to the stock 2050 out there (if not better).
Ngl, with the expectations that I had from the Steam Deck, my hopes actually raised a bit. This is a lower-end expectation, and it's already kicking the Steam Deck's ass. Comparing Control on the Steam Deck to Control on the underclocked 2050 made me cry tears of joy.
If Nintendo made a remake or sequel to Eternal Darkness, they would make the most legitimately unnerving game ever made. Imagine the game just slowly dissolving as the sanity meter goes down. I'm thinking something like a merger between The Void from the Binding of Isaac and The Clockwork Mansion from Dishonored 2. That shit would be insanely cool.
Yeah, that's insane. It's already kicking Steam Deck's ass on the expected handheld mode, even at native 720p... An off-the-shell chip we and the folks at DF are using to approximate the actual T239, I don't get how someone can look at this and somehow lower their expectations. Heck, it's already a few frames behind Rog Ally in some of these tests (while handheld!), and don't even get me started on the RT performance... I'm very excited man, Switch 2 will kick ass.
It was 750 mhz, also keep in mind it's 16sm
People are already aware of that, 4 more SM isn't enough of a difference to skew these results when using Windows and unoptimized ports are the samples.
Procedural generation isn't AI driven though. It's probably best explained by how Minecraft or the Binding of Isaac does it. Essentially, there's a mapped set of combinations that a game can have, and a "seed" can generate X amount of combinations for a world. The longer the seed, the more combinations that can generate.
My point was that clocks were higher than at least I expect in portable mode,+ more silicon.
That might be what you expect but 750mhz is very much on the table. It's a very small difference that can even be upped by specific games like with the Switch (assuming we're talking about Thraktor's 660mhz as the base).
I didn't do any comparisons to the ROG Ally due to inexperience, but the fact that's the case... and will likely cost much less? Goddamn we've got a cool as hell system on our hands. Genuinely can't wait.
As an intentional style that kinda static haze could actually work for a game.
You can find benchmarks with the Ally running these games, the biggest difference shows on Cyberpunk by a long shot.
Thraktors analysis was 550 MHz?
It was? Might be mixing them up, I do know NVN2 pretty much stated 660mhz (while on 8nm, to note).
