One of the big question marks hanging over the Nintendo Switch is just how powerful it can actually be. As a hybrid console that has to straddle the divide between home and portable gaming, it’s a fascinating prospect, but that pitch in and of itself demands that certain compromises be made.
Would you believe it, but details of the console leaked yet again late last week, with VentureBeat confirming some of the previous suspicions surrounding the console’s chipset, which Digital Foundry have further analysed today alongside their own sources. In particular, it’s that the chipset is a modified version of Nvidia’s Tegra X1, with the GPU being based on their last generation Maxwell GPU design, as opposed to Pascal which is to appear in the Tegra X2. There is a degree of common ground between the two, but how much of Pascal has filtered through, if anything, isn’t yet known.
What is now fairly certain is that the chip is dramatically under-clocked compared to a standard X1. The CPU’s eight cores now top out at 1020Mhz, roughly half of the standard X1, while the GPU drops to just 40% of the designed speed when in portable mode. To be fair, Digital Foundry speculate that this is more than enough to match or better the Wii U, and has been done specifically to meet Nintendo’s targets for performance and battery life.
That said, when docked the Switch GPU steps up to full capacity, offering 2.5x the performance which should be able to handle the jump up to 1080p. In some ways it parallels the way in which developers now have to consider the PS4 and PS4 Pro, or the Xbox One and upcoming Scorpio, with the same game running in two different configurations.
In either configuration it won’t be as powerful as the PlayStation 4 or Xbox One, but Nintendo have always had a knack for getting the very best out of their own hardware, so I’m very eager to see what’s in store when the console is shown publicly in mid-January.