Soundwave said:
Manlytears said: I doubt it. I agree on the GPU, it is quite possible that Switch 2 can deliver a "minimally comparable" GPU, but what about the CPU !? I find it very difficult for nintendo to make a Console/Handheld with GPU and CPU comparable to the "series S" rumored specs and, at the same time, maintain portability and low price (both characteristics of Nintendo's most successful consoles) |
Arm A78 in a 8 core config isn't actually that far off from the CPU inside the XBox Series S.
GPU performance is far more important for ports.
|
nVidia typically builds it's own CPU cores. Aka Denver. And thus does not generally rely on already established Cortex designs.
Soundwave said:
High speed storage isn't magically exclusive to consoles (no matter how much Sony kool-aid you drink). There are already mobile devices like iPhones that have had NVMe storage for years now. Android phones have NFS 3.0, but in a few years even faster NFS 4.0 will be available which will be as fast or faster than the XBox's drive. The world doesn't revolve around game consoles.
|
Don't equate me with someone who drinks "console advertising kool-aid". I'm a PC gamer primarily.
I think I have demonstrated over the years on this forum to have a slight understanding of technology, don't you think?
The iPhones are NOT running NVMe drives like the Samsung 970 Evo with 4x PCI-E lanes, 2GB of built in DRAM... That drive alone will happily consume 6w or more of power at load... Do you honestly think a handheld can sustain that? No. No they can not.
And that isn't even the fastest and thus most power hungry drive on the market either... Food for thought.
Apple took the Macbook SSD nvme controller and modified it for the iPhone, it's still not the same as a desktop or laptop nvme drive... And it is certainly far removed from the next-gen consoles which have employed extra silicon to aid in the compression and accelerate transfers.
Rather Apple has taken PCI-E over a MIPI M-PHY physical layer.
MIPI M-PHY is the definer of the physical characteristics of the interface, where-as PCI-E is what we do with the channels.
In short... A mobile device is physically limited, so you cannot use the highly parallel nature of memory transactions to increase bandwidth exponentially... Thus a Switch 2 will always be behind something more forward thinking.
Soundwave said:
Ray tracing ... please go ahead MS and Sony, pretty please. Please do push Ray tracing. It will cripple both the lukewarm RDNA 1.5 GPUs in those machines (and a no-go on Series S entirely) in performance faster than you can say "I want a Nvidia 3080". The PS5 in particular will struggle with even PS4+ range graphics when you start introducing legit ray tracing elements into a game.
|
The Playstation 4 was already having games leveraging Ray Tracing. It's not a new concept.
But hardware Ray Tracing will enable all sorts of things going forward.
Developers will come to grasp it's various nuances and leverage that substantial amount of compute hardware effectively over time, no doubt about it... As for performance, it's to early to speculate how extensive AMD's implementation is.
Soundwave said:
DLSS is a game changer for performance, the simple fact of the matter is it's much easier for any GPU to run a game at say 540p than 1800p or 4K. That's not some magical rule that applies only to Nintendo is specialized situations. There just was no great way in the past to be able to get the nice picture quality from higher resolutions without actually rendering the higher resolutions. DLSS changes that.
|
I never argued that DLSS wasn't beneficial.
However, it's effectiveness does vary game to game, scene to scene.
Personally? I am actually happy with 1440P with the visual settings dialed up on PC and pushing 144hz instead... And if I have the hardware overhead, you bet I will super sample a 2160P image.
However... You also need to remember that the next-gen consoles do have their own alternatives to DLSS... The Xbox Series X/S for example will be leveraging Direct ML, which is Microsofts alternative, so there is absolutely ZERO reason why the Xbox Series S/X cannot also render a game at 540P and dial up the settings even further still. (Unlikely, but it's a scenario.)
https://docs.microsoft.com/en-us/windows/win32/direct3d12/dml-intro
Soundwave said:
Bench pressing 150 pounds will always be easier than benching 250. That's basically what DLSS does for a GPU it lets it give the appearance of "benching 250" when in fact the processor is really only lifting 100 pounds.
|
Silly comparison in my opinion, I understand what DLSS does and it's implications.
Manlytears said:
I am not an ARM expert, but I highly doubt that a possible Arm A78 running on a portable console can reach a "minimally comparable" level of 3.8ghzs Zen2 based CPU that has unlimited access to electricity ... Even if they made "Switch 2" with only 1-2h of battery life, something that really isn't going to happen, it wouldn't be possible.
|
Not to mention Zen2 is possibly one of the best CPU designs out currently...
Soundwave said:
Wait till you see the new Apple chips that are coming, there's a reason why they're ditching Intel for straight up mobile components even in their Pro laptop and desktop lines.
Mobile CPUs are no joke. There's also the ARM X1 core that is coming that could be a possibility for Switch 2 if Nintendo really wanted to go pound for pound, but I don't think honestly despite all the hype that CPU's will be maxxed out by games anytime soon. Game development is way more focused on GPU for a long time and it will remain that way because many devs simply don't want to think outside of the box. That and there's still a crap ton of PC gamers with old ass legacy Intel CPUs that they want to support.
|
And Apple reduced the prices of the ARM devices too. Why? Because the ARM chips are lower performing than AMD or Intels best. (Although more energy efficient.)
Apple can also reduce it's cost overheads as it cuts out the middleman, very good value proposition for a company that loves money.
https://www.extremetech.com/computing/312234-apple-a12z-arm-performance-vs-x86
JRPGfan said:
*edit2: "If MS had a Series S equivalent for the current XBox One, it would've been basically 400 gigaflops ... which is the same as the Switch docked."
Thats probably true... 1/3th the power, of a Xbox One, is roughly a Switch. 1300/3 = 433 Gflops.
|
Except the Xbox One @300Gflops would be faster than the Switch.
It would still have 8x 1.75Ghz Jaguar cores which beats the Switch.
It would still have the ESRAM to boost memory transactions.
It would still have 8GB of DDR3 memory on a 256bit interface @68.3GB/s.
I thought this forum had finally moved from flops as a determiner for performance...
If everything was kept equal, same CPU cores, Memory etc' and the only difference was the GPU's, the Switch would win hands down, Maxwell is a far more efficient GPU architecture than Graphics Core Next 1.0.
Soundwave said:
AMD has not shown a direct comparable to DLSS 2.0 as of now. Microsoft has a machine learning solution but they have also not been forthcoming in showing it working really. They had a panel in 2019 GDC but have not mentioned it again at all and not in any XBox Series X discussion whatsoever which I find very odd if this is supposedly a feature its supposed to have.
|
AMD's main counter to DLSS is Radeon Image Sharpening. Not the same. But does a good job.