The Nintendo “Switch” was announced this morning, the next-generation half-portable, half-docked console. To reduce confusion, the Switch was previously referred to as the Nintendo “NX.” It is the same device.

Nintendo's new Switch is built in partnership with nVidia and leverages the Pascal architecture found in current-generation GTX 10-series GPUs. At least, based on this text from nVidia's blog: "[...] NVIDIA GPU based on the same architecture as the world’s top-performing GeForce gaming graphics cards." Tegra SOCs include ARM processors alongside the nVidia graphics solution, and also host all of the I/O lanes and memory interfaces. This is a complete system, as indicated by “system on chip.” We've asked nVidia for details on which ARM devices are used and which memory will be supported, but were told that the company is not revealing further details on Nintendo's product. We are awaiting comment from Nintendo for more information.

We do know that the Tegra SOC is accelerating gameplay with hardware-acceleration for video playback, and that nVidia and Nintendo have deployed “custom software for audio effects and rendering.” We can confidently speculate that the Switch is not functioning as the previous Shield devices have (read: not streaming to handheld from a dock), mostly because the Switch is large enough to contain all necessary render hardware within its handheld state. The Switch is also shown in the advert to be playable on planes, which most certainly do not have fast enough internet to support up/down game streaming. This is processing and rendering locally.

Unreal Engine 4 -- one of the premiere engines used for creating games -- has been making quite a splash in the gaming market, primarily due to demonstrations that show off its impressive potential. More recently, Epic unveiled an Unreal Engine demo using nVidia’s Tegra K1 mobile SOC, which hosts a 64-bit ARM CPU and Kepler-based GPU with 192 CUDA cores, 4 ROPs, and 8 texture units.

After offering reddit's computer hardware & buildapc sub-reddits the opportunity to ask us about our nVidia GTC keynote coverage, an astute reader ("asome132") noticed that the new Pascal roadmap had a key change: Maxwell's "unified virtual memory" line-item had been replaced with a very simple, vague "DirectX 12" item. We investigated the change while at GTC, speaking to a couple of CUDA programmers and Maxwell architecture experts; I sent GN's own CUDA programmer and 30+ year programming veteran, Jim Vincent, to ask nVidia engineers about the change in the slide deck. Below includes the official stance along with our between-the-lines interpretation and analysis.

In this article, we'll look at the disappearance of "Unified Virtual Memory" from nVidia's roadmap, discuss an ARM/nVidia future that challenges existing platforms, and look at NVLink's intentions and compatible platforms.

(This article has significant contributions from GN Staff Writer & CUDA programmer Jim Vincent).