Competition is getting very stiff. There was a time when the only mobile SoC outside Apple Land that could be considered quite decent was created by Qualcomm. Now we have Samsung who has released Exynos which will soon have LTE in the 5430 iteration. MTK has come with quite a few very good bang-for-the-buck SoC and has LTE MT6595, MT6752 and MT6732 right around the corner.
Entering the arena with a serious SoC is juggernaut Nvidia who previously put out Tegra 3 and Tegra 4. Now they’ve decided it’s time to enter the battle full throttle with the Nvidia K1.
| Tegra K1 | |
| GPU | |
| NVIDIA® Kepler™ Architecture | 192 NVIDIA CUDA® Cores |
| CPU | |
| CPU Cores and Architecture | NVIDIA 4-Plus-1™ Quad-Core ARM Cortex-A15 “r3” |
| Max Clock Speed | 2.3 GHz |
| Memory | |
| Memory Type | DDR3L and LPDDR3 |
| Max Memory Size | 8 GB (with 40-bit address extension) |
| Display | |
| LCD | 3840×2160 |
| HDMI | 4K (UltraHD, 4096×2160) |
| Package | |
| Package Size/Type | 23×23 FCBGA 16×16 S-FCCSP 15×15 FC PoP |
| Process | 28 nm |
What stands out at the top of the chart is the 4-PLUS-1 technology. Previously this tech was known as ‘variable symmetric multiprocessing’, but Nvidia fans wanting a bit more marketing friendly term were eventually treated to the easier on palate and mind — ‘4-PLUS-1’. This is a similar concept to big.LITTLE, but instead of 4×4 or 2×2, it’s 4×1, where 4 cores kick in to handle the heavy loads and 1 core is used when the load is light.
The CPU to be used will be ARM Cortex-A15 with a max clock speed of 2.3GHz.
Very interesting to note the Tegra K1 is speccd to push up to 3840×2160 resolution and will come stock capable of outputting HDMI at up to 4K UltraHD (4096×2160). We see right away this SoC will be an excellent candidate for large tablets wanting to have high-end screens capable of retina displays. On an 11.0″ tablet screen, 3840×2160 will give you a PPI of 400. At this PPI, it will not be possible to discern individual pixels.
From the official Nvidia site:
 |
SMX Delivers more processing performance and efficiency through this new, innovative streaming multiprocessor design that allows a greater percentage of space to be applied to processing cores versus control logic |
 |
Dynamic Parallelism Simplifies GPU programming by allowing programmers to easily accelerate all parallel nested loops – resulting in a GPU dynamically spawning new threads on its own without going back to the CPU |
 |
Hyper-Q Slashes CPU idle time by allowing multiple CPU cores to simultaneously utilize a single Kepler GPU, dramatically advancing programmability and efficiency |
The Nvidia K1 will use Keplar Technology. This is the same tech used in their PC GPUs. While it’s likely not as power-saving as Nvidia claims, they have been in the GPU game for about as long as modern 3D gaming has been around, so we’re expecting excellent performance ratios.
Let’s see how Tegra K1 fairs against the competition. The units used for the testing is Tegra 4 Nvidia Shield, Snapdragon 801 Oppo Find 7 and a Tegra K1 unidentified unit with 4GB RAM and PPI of 320.
What a thrashing. Though resolution plays a large part in test scores, we must assume the K1 at the very least was competing with a 720p device.
It’s enough to convince Xiaomi, who has contracted the Nvidia K1 to be used in their upcoming MiPad tablet line.
A year or so back it may have been MTK who had some of the Big Boys shaking in their boots, but with the super hot SoC race and plummeting prices all around the block it is now MTK who really needs to step up their game.
It’s now possible to buy near premium phones such as the IUNI U2 with Snapdragon 800 in the $300 ballpark. While 2013 saw us deciding between a $200 MTK based phone and an $800 iPhone or Samsung Galaxy, the gap has lessened and many consumers looking to spend $200-250 may decide to spend 30% more for the SoC brand recognition alone.
Leave a Reply