First, though, a quick note about the test setup: NVIDIA had a handful of octa-core X1 reference boards with 4GB of RAM and tablet-class heatsinks slapped onto them (an NVIDIA spokesperson said the setup was essentially what you'd find in an 8-inch X1-powered tablet). As Huang noted on stage, the X1 chipset also packs the 256-core Maxwell GPU; that's up from the 192 last year's K1 came loaded with. Alas, NVIDIA wouldn't confirm how fast the processor cores themselves were clocked. C'est la vie, friends.
Anyway, take a peek at the numbers the X1 put up. Where applicable, we've included scores from the Nexus 9 and the NVIDIA Shield tablet, which run the 64-bit and 32-bit versions of the Tegra K1, respectively.
See those two AndEBench values in italics? We didn't have a Nexus 9 or a Shield Tablet on hand to test, so we used median values provided by the app's maker, the Embedded Microprocessor Benchmark Consortium. The X1 pulls ahead of both its predecessors in the test, if only just (we'll update this post after we run our own tests on the proper hardware). The X1 is no slouch at video, as it supports 4K video decoding at 60 frames per second -- it handled playback of a video shot from a Red camera without a single hiccup, just in time for all the big content providers to, well, go big on ultra high-def video.
Maxwell's power is undeniable, too: The X1 absolutely spanked all comers (sometimes by over a factor of two) in the more graphically involved tests, a testament to the strength of its future as a game platform. While we're on the subject of gaming, NVIDIA also confirmed that the X1 would play nice with a whole host of APIs used in game development, since it has the same Maxwell architecture as its high-end GTX 980 desktop graphics card. In a perfect world, that'll make for easier ports and less hectic game translations from more traditional platforms like PCs and consoles to mobile ones.
And the icing on this slice of silicon cake? The X1 is designed to sip (rather than gulp) your device's power reserves. In one particularly involved test, the NVIDIA team measured the voltage usage rates of the X1 versus the iPad Air 2... after tearing apart eight of them and downclocking the X1 so both were at an equivalent level of performance. In the end, average power consumption for the Air 2 was 2.6 watts, versus 1.5 watts for the X1, a pretty significant power savings. It's still early days for the X1, and all of the above has been just a taste of what we can expect when X1-powered devices hit later this year.
Nicole Lee contributed to this report.