A new series of Intel's Atom processors is here ostensibly for IoT applications. The company claims the E3900 series is designed from the 'ground up' so that we have a processor that is exceptionally capable of delivering on performance, processing and scalability.

Intel also claims new levels of security, determinism and image and video processing power for the MPU. And there is a variant specifically intended for automotive applications: the A3900 series. The "A" for automotive version is targeted at software-defined cockpit solutions like in-vehicle infotainment (IVI), digital instrument clusters and ADAS -- all of it based in a single SoC. Intel says the A3900 series will allow car manufacturers to offer new levels of determinism for real-time decision-making. It is sampling with customers and will be available in Q1 2017.

The Intel Atom processor E3900 series will make the edge and fog [computing] more intelligent – enabling many of the processing needs to take place at or near the data sensor and alleviating the need to push all processing to the data centre, according to Ken Caviasca, VP at Intel's IoT Group and its GM of platform engineering and development.

This series delivers more than 1.7x more computing power over the previous version. Intel has aimed for faster memory speeds and memory bandwidth to improve processing capability needed for edge to cloud network computing.

The 14nm node chip is packaged in a flip chip ball grid array (FCBGA). Additional new features and capabilities include:

  • Intel's ninth-generation graphics engine to improve 3-D graphics performance by 2.9x over the previous generation and support as many as three independent displays.
  • Four vector image processing units that result in better visibility, quality video in low light, noise reduction and colour and detail preservation.
  • Intel Time Coordinated Computing Technology that coordinates and synchronises peripherals and networks of connected devices. By synchronizing clocks inside the SoC and across the network, you can achieve network accuracy to within a microsecond.