Was this a recent development? I ask because am I being an eternal optimist in wondering if one of the unrevealed special things about switch is that it will allow you to stream your PC games to it?
It could be a good tick on the box, and because Nvidia has already done it, the partnership could mean this gets moved to Switch.
As JoshuaJSlone mentioned, they officially confirmed that they were dropping the Shield tablets recently, but the decision was probably taken a while ago (there was a TX1-based Shield tablet expected to hit earlier this year which mysteriously disappeared).
I would be extremely, extremely surprised if you could stream your PC games to Switch. Nintendo's entire business model is based on the fact that you can only buy software for the system through them, so allowing people to play games purchased elsewhere (even if it is only streaming) certainly doesn't sound like the kind of business decision they'd make.
Ah, ok. Well, I guess it's certainly plausible then. I wonder how they would approximate the additional threads in the dev kits then? I suppose they could have told developers to plan for them, but it would be suboptimal.
I suppose it depends on the final configuration. If it's 4x A72 + 4x A53, then it should be able to handle anything built for the X1 without issue, and the worst case scenario would be developers not making full use of the A53s. If it's 2x A72 + 4x A53, then it gets a little messier, as devs might have issues adequately parallelising across the four lesser performing A53s.
In any case I'm sure they would indicate what the final configuration and expected performance is (keeping in mind that clock speeds can vary until very late in the game anyway), and I think it's just one of those things you have to deal with when developing for unfinished hardware.
Well it is less about Tegra and more about Nvidia as a whole. The shield itself was suppose to be a showcase of their tech so that other people could partner up and use the tech. Part of the issue was they didnt have the software library to grow the idea... The Switch has a software library and a set of Nintendo developers who are wizards.
This is literally a tech showcase for hardware and software for Nvidia. If you take Tegra out of the picture it is still using tech from their desktop cards. This still could push things like Gameworks, GRID, Shadow Play, and Emulation.. This is a trojan horse to developers who decide to make games for the switch.. it could push Nvidia middleware onto them.. So I think this is a way for Nvidia to market themselves if anything
The Shield was a showcase of their Tegra tech, precisely to get Nvidia the kind of big client they've just got with Nintendo (hence why they're dropping the Shield line as soon as they've scored that client). A big, high-volume client like Nintendo (with opportunities for long-term recurring business) was the end-goal of the Tegra division, not a step along the way to something else.
Nintendo aren't going to use Shadow Play or GRID, and I don't see any reason for them to outsource VC emulation to Nvidia, so Switch wouldn't help push any of those things for Nvidia. Third parties will probably get access to versions of many of the GameWorks libraries on Switch, but I don't know if it would lead to much of a push in the PC space (the market for third-party games on PCs with a Nvidia GPU is likely far larger than the market for third-party games on Switch will be, so if they're not using GameWorks now, I don't see Switch presenting a huge pressure for them to change their minds).
There would potentially be some benefit to developers optimising for Nvidia's architecture on Switch and bringing those optimisations over to PC, as there's good reason to believe that this is happening for AMD as a result of PS4 and XBO. However, unlike PS4/XBO for AMD, Switch isn't suddenly going to account for the vast majority of Nvidia powered hardware out there, and it's certainly not going to become the default platform for AA western games. Some big third party engine devs will put the effort in to optimise for Switch (because their job is to provide an engine that's runs well on every platform its available for), but they're likely to be the engines, such as UE4, which are already pretty well optimised for Nvidia PC hardware anyway. Small indies who focus on Nintendo hardware will likely put in the effort too, but if their games do come to PC they're unlikely to require anything but the most entry-level hardware. Lastly are the mid-tier engines, which will possibly be lightly optimised for Switch, but probably not to the extent that would translate to big improvements to Nvidia PC performance (in many of these cases I can see developers putting their work into figuring out which parts of their shader code they can switch to FP16, which would be of no benefit to Nvidia's PC hardware, but perversely
would be of benefit to newer AMD cards).
That's not to say I don't think that Nvidia at large wouldn't stand to benefit to some degree from having their hardware in Switch. A few extra people in the world working to optimise games for Nvidia's graphics architecture certainly wouldn't hurt them. I don't however, think that this benefit would be anywhere near enough for them to take a hit in a deal which is worth potentially billions of dollars in long-term revenue to them.
No, i'm pretty sure i don't have this backwards. I also have no problems with having the guy from Nvidia hype up the device during the conference and control the presentation. Like i already said.
But I'm talking about very specific specs. I think you're gravely mistaking if you think that the average mainstream journalist is going to be looking into which Arm cores at which fab node are being put inside the system if the press release with this info is coming from Nvidia at a different time/place. He/she might look into it though when that info is disclosed during the main event. So they can hype it up for the average Joe with "Now you're playing with power" and "Blast Processing" slogans, present it as the most powerful portable device ever... no problem, but i'm talking about the actual specs.
I'm talking about actual specs. I'm talking about press seeing numbers like 750 Gflops and 4GB of RAM in a press release and saying "they're lower than the numbers Sony and MS have, so this must be a pretty weak machine". Presenting stuff like that well, in a positive context on stage can make a
huge difference to how it's interpreted compared to just dumping it on a press release.
To give a stupid off the top of my head example, consider the following two options for presenting the same piece of information:
Press Release said:
Floating-point performance: 758 Gflops (FP32) / 1516 Gflops (FP16)
Stage Presentation said:
In order to achieve Switch's unique portability without compromising on performance, we've optimised the N1 processor for ultra-efficient FP16 computation, which runs at twice the speed of traditional GPU code and consumes far less power. We've managed to squeeze in over one and a half teraflops of performance with this highly efficient technique, providing true console quality gaming while still allowing players to play however, and wherever they want. We've also been working with developers to help them make the most out of Switch's immense FP16 capabilities, and we've been amazed at what they've managed to do, with even launch games like Pikmin 4: The Revenge of The Pikmin managing to run almost 80% of their graphics code in this high-performance mode.
Anyone reading the first will simply compare the numbers to what Sony and MS has published, see that their FP32 figures are much higher and conclude that Switch is well behind the competition and that FP16 isn't really used in games so doesn't count.
Someone watching the second, though, will get the impression that FP16 is super-efficient secret sauce that Nvidia invented to allow Nintendo to cram console hardware into a portable, and that maybe it is possible that this really competes with XBO and PS4 in such a small box. It even pre-emptively counters the claim that games can't really make use of FP16 by presenting one game that does. Never mind that that game happens to be a complete outlier and most third parties are closer to 20% utilisation, as people are going to latch onto the one data point they're given, not ponder the possible distribution. Never mind, either, that it might only achieve half or less of that 1.5TF of FP16 when in portable mode, because most people will simply never think that hard about it.
All it takes is the right game to pull that crowd onto the platform. I have said this many times..If Nintendo worked (or is working) with Psyonix to get Rocket League on the Switch It would be the trojan horse Nintendo needs to drag a portion of core gamers onto the platform.
Rocket League in my eyes is tailor made for the Switch. Local multiplayer in conjunction with online play... specific crates and cars based around Nintendo IP and so on..
Specs wont matter that much if you can generate enough buzz and bring the right titles to build the platform. No one in their right mind would have thought pokemon go would be a thing.. hell even the developers didnt anticipate the culture explosion.
I'd definitely be willing to double-dip with Rocket League on Switch. Remote play on Vita is no substitute whatsoever for being able to play it directly on a lag-free portable.