| bonzobanana said: That video doesn't really factor in battery power to their equation but their figure for portable performance is comparable to GTX 750 Ti which is about 1.4 Teraflops but consuming up to 12-17W, someone has put a figure of 800 Gflops as peak performance in portable mode for 2 hours use. I assume they have done the analysis of battery consumption. On face value I would of thought my figure of 600 Gflops was more accurate as it can probably only consume a max of 5-6W in total excluding the screen. So the Switch 2 can do short life peak power of around 1.4 Teraflops in portable mode but it can't maintain that for 2 hours. Of course the final figure is not known, maybe Nintendo have over-stated the minimum 2 hours runtime and the reality is only 90 minutes for portable mode and the clocks can go up again. Everyone is agreed that 20Wh is the battery though and this video confirms a much more power hungry chipset than first thought. I guess we need someone to analyse the power consumption of this chip based on this new information from this video. We are still guessing with regard portable performance because the 20Wh battery doesn't seem to be factored in. There is going to be fantastic analysis in a few weeks I'm sure of retail hardware and how much power it is consuming in portable mode. Ultimately I think we have to be realistic based on the information that is coming our way at the moment, the chipset is a much cheaper fabrication process than first expected and it is clear the design is relying on DLSS upscaling heavily to compensate for a cheap power hungry chipset. The picture is building of a fairly low spec console with regard graphics hardware but very decent CPU performance that takes it way beyond PS4 for example. It's a clever design that has enough CPU performance to do the main code of modern games with optimisation and then the output state of the graphics is upscaled. For a long time the information we have heard about Switch 2 was a console that relied heavily on upscaling and that is the reality surely that we know now. |
The reason why you can't compare the power-efficiency of a (down-clocked) RTX 2050 to the power-efficiency of a T239 is because they are different chips that target different power profiles, have different SOC (not micro, SOC) architectures, and therefore different power-efficiency curves. The comparison makes sense from a performance perspective (even if rough) but not the power-efficiency perspective. There also is the matter of the different CPUs.
This is true even if the GPUs are both Ampere.
I'll provide an example of why this is the case,
Consider the RTX A2000 and RTX 3050. They both have a GA106 die. The A2000, however; has 3328 cores clocked at 562* - 1200 Mhz while the RTX 3050 has 2560 cores clocked at 1552 - 1777 Mhz. The A2000 slightly performs better than the RTX 3050 in games and compute. Yet the RTX A2000 pulls about 70W at full utilization and the RTX 3050 pulls about 115W - 130W at full utilization in its original form.
Now this is for cards that have the same GA106 (just cut and clocked differently.)
The GA107 in the RTX 2050 Mobile is even more different from the T239 than that.
*Notice the 562Mhz here, this likely isn't a coincidence, that probably corresponds with an actual power-state.
As to the exercise of halving the TFLOPs, you'll notice that the 1.84 Tflops PS4-equivalent performs worse than the 1.39 TFlops GTX 750ti on that benchmark. TFlops and real-world performance are only loosely related, and you really can't extrapolate one from the other when comparing the performance across architectures.
Also the curves are not linear, so halving the GPU's power utilization doesn't necessarily mean halving its performance. The curves do become near-linear after some low-point, but there is reason to believe that a T239 has that point shifted leftward compared to an RTX 2050, due to their different power-targets.
As an example, this is what a power-performance curve of an A5000 (light-blue) looks like.
or the power-performance curve of a Steam Deck Oled.
Notice that at 6.5W it is retaining 83% of the performance as it has at 12W (1500Mhz vs. 1800Mhz, all else equal.) Power utilization nearly halved (6.5/12 ~ 55%), but theoretical performance** didn't (1500/1800 ~ 83%.)
**TFlops are proportional to clock-rate and core-count. Core-count is a constant here.
Last edited by sc94597 - on 10 May 2025
