NVIDIA unveiled the processor Tegra K1, developed for mobile devices, a year ago at CES 2014.
Tegra K1 is a system-on-a-chip (SoC) processor that integrates all components of an electronic system into a single chip.
According to the press-release, the first version of Tegra K1 uses a 32-bit quad-core A15 CPU (central processing unit). The second version uses a custom, NVIDIA-designed 64-bit dual-core CPU codenamed Denver.
Denver-based Tegra K1 processor
Denver CPU has a 7-wide superscalar architecture, meaning it can execute seven instructions per clock cycle compared with 3 instructions per clock with the A15. NVIDIA has stated that their dual core Denver CPU can surpass quad- and octa-core mobile processors on most mobility workloads while delivering great power efficiency.
NVIDIA has also provided the public benchmarks results for the Denver-based Tegra K1 against current high-end mobile processors: Haswell Core-based Celeron 2955U used in many Chromebooks, Apple’s A7 Cyclone used in the iPhone 5s, Krait-400 (8974-AA) and Bay Trail (Celeron N2910).
The Denver processor “significantly outperformed” the ARM-based and Bay Trail processors on all of the benchmarks, and it delivered a similar level of performance to the 1.4GHz dual-core Haswell processor.
NVIDIA claims that Tegra K1 processor delivers stunning graphics and visual computing capabilities through the 192-core GPU (graphics processing unit) featuring Kepler HPC (high performance computing) architecture.
About NVIDIA Kepler architecture used in Tegra K1
Previous Fermi architecture used Streaming Multiprocessor (SM) unit, which contained 32 CUDA* processing cores. Current Kepler architecture uses next-generation Streaming Multiprocessor (SMX), which contains 192 CUDA cores. CUDA cores on the SMX run at a lower clock speed than on the SM, which allows the SMX to use less power and deliver more performance at the same time.
Originally the Kepler architecture was designed for notebooks and desktops, later it was added to workstations and supercomputers, and now it’s available for smartphones.
* Official NVIDIA blog describes CUDA as a parallel computing platform and programming model that makes using a GPU for general purpose computing simple and elegant.
Mobile devices based on Tegra K1 processor
Currently Tegra K1 processor is used in Google Nexus 9 and a number of Chromebooks by Acer and HP. Xiaomi MiPad uses this CPU as well.