The fans running in portable mode certainly is strange if these clocks are accurate. I mean even at the docked rate you likely could get by without fans if there are only 2 SM.
Perhaps these clock rates aren't max rates for portable and docked but minimum?
And if devs need to they can increase rates to a max rate that EG does not have information on?
Otherwise yeah best case we are hoping for 3-6 SM
There shouldn't be any maximum or minimum clock speeds, the CPU and GPU clocks will be constant (while running games anyway, the OS is a different matter).
I wouldn't even consider 6 SMs as a best case. Consider it a "better than best" case.
I'd be kinda shocked if a fan needs to run in portable mode. That sounds rather noisy for a handheld gaming device. Plus, more moving parts would make the Switch more fragile than handheld systems in the past.
I mean, do any phones or tablets out there need a fan? (I'm genuinely asking).
Occam's razor says that among competing hypotheses, the one with the fewest assumptions should be selected. I'm just saying...
No phones, but some tablets do. This is a little different in that it's designed both for stationary and portable use, though.
Thanks for the write (I'll pretend I understood what you wrote, but it does sound more positive than what others have been saying).
Two questions:
1. How much higher than Wii U are we talking here regrading each scenario? Most people are fine with the portable aspect, but it's the home console aspect that is making people weary. How does it compare to XB1 and what should we expect from ports to Switch? Is it merely less graphical fidelity or will it be something more severe?
2. It seems that active cooling is what's not making any sense for the Switch, but then I wonder, what is the possibility of it simply being badly designed? It might seem like crazy talk given that it's Nintendo we are talking about there. But this is really the simplest reason for it.
For question 1, it's difficult to say, as it depends on whether the game is CPU or GPU limited, how well it can use FP16, etc. In terms of third party ports, I'd expect 720p with some effects disabled when docked, and 540p in portable, but that's pretty much what my expectation's been for a while.
On question 2, there's not much "bad design" that Nintendo can do here. TSMC physically make the chips, and Nvidia's GPU architecture is pretty power efficient (certainly compared to AMD). Nintendo isn't really in a position to screw either of them up. It's possible that they were just extra-conservative with heat dissipation when including a fan, but they're usually extra-conservative with moving parts, so I don't feel they'd use a fan unless they really needed to.
Excellent analysis as always, and I agree with that conclusion regarding the fan, though I'm starting to expect that I'll be disappointed again regardless of what I'm hoping for.
But another thing to consider is LKD's report that there is an additional fan in the dock. I know the patent didn't support that report but the patent was filed in June, and adding a fan to the dock is certainly a possible change from the patented specification. 2 fans would be a ridiculous amount of overkill if the clock rates listed are being ascribed to TX1 hardware.
Also, consider the DF article did mention that it's likely that some of the customizations to the SoC were from the Pascal architecture, so it's still certainly possible this will be on a 16nm process. Unlikely I'd say at this point, but possible.
I'm pretty much of the belief that the fan isn't being used in portable mode at this point, as they would have to be using a much bigger GPU than I'd expect to warrant it at these clocks. It is puzzling, though, that as of June they were expecting to have to run the fan in portable mode, as they would have had to know the final GPU config at that point, and even assuming they reduced the clocks since then for some reason, they still shouldn't have needed it at the time.
At a guess the extra fan in the dock may just be for non-final dev-kits with early hardware the runs hotter (although again it leaves us the question of why the power draw would have gone down), but I doubt there's one in the final hardware.
The article actually explicitly left open the possibility of this being a 16nm chip. I'm not expecting it, though, as you could probably double these clocks on 16nm and still be fine (unless we're talking about an enormous GPU).
So regarding the possibility of more SMs:
I believe according to graphs posted here, you get higher performance per watt with more SMs than you do by simply increasing the clock speed. Assuming Nintendo is targeting something like 5 hours of battery life they would be better off using a lower clock speed regardless of the amount of SMs.
So when we look at this from a design perspective, the two cost related variables here are:
- Die cost/SM cost
- Battery cost
And two power related variables here which directly affect the cost variables above:
- # of SMs/CUDA cores
- Clock of SMs/CUDA cores
It may be possible that Nintendo opted to increase the number of CUDA cores to reach their target performance rather than increase the clock speeds, as increasing the clock speed would require a larger and more expensive battery. Does anyone know how much money a larger battery would cost a console maker relative to an additional SM on the die?
Yeah, that's basically the trade-off you make between more ALUs at a lower clock versus fewer at a higher clock (it's the same for CPU cores in arbitrarily parallelizable tasks). If Nintendo has gone with more than 2 SMs, then it would seem that it's to get similar performance at a bit lower power draw rather than to push the performance ceiling up.
Regarding batteries, they're typically limited by size and weight more than cost. There's going to be a certain amount of free space inside Switch to be filled by battery, and if Nintendo wants more they have to increase both the size and the weight of the device, which is likely a bigger deal than the cost of the battery itself.
Unfortunately, multiple people are saying that the leaked specs are correct other than clock speeds. It's 2SMs. 3 SMs matches the rumored reports of its performance perfectly, as I said earlier in the thread, but that just doesn't seem to be the case. If it were 3SMs I'd be extremely satisfied with these clocks, but it just isn't.
I don't understand it, though! 2SMs just doesn't make sense unless NVAPI is the greatest graphics API in history or devs have quickly taken to FP16. It just doesn't make sense in my head and I can't reconcile it.
I don't think we've had a single reliable report that has told us the final GPU configuration (a number have recited the TX1 specs, but nothing more), and even in this article they specifically say that they don't know the number of "Cuda cores" the GPU uses. The "leaked specs" are simply a recitation of the TX1 specs from someone who a week prior had insisted that Nintendo weren't using Nvidia, and we know they've got both CPU and GPU clocks very wrong, so I don't know why we're putting much stock in them.
