Technical Articles

3 ways All Programmable SoCs create opportunities | Avnet Silica

Display portlet menu

3 ways All Programmable SoCs create new opportunities for designers

Two men and woman reviewing technology products.

Usually, two heads are better than one. Yet sometimes, they introduce conflicting perspectives and priorities.

Two heads means two places need processing power to handle data. Two heads with varied performance have two different effects on the final product. Plus, two heads means two new places with constantly changing technology, which hits software designers especially hard.

In the growing market of embedded vision, there’s a need to not only rapidly scale to compete but to also be ready for machine learning’s effect on the space. That’s all while keeping up with lightning fast design cycles. Luckily, that’s also where FPGAs merge two heads into one: all the processing power is happening on hardware. Instead of needing an external operating system, FPGA rewires the chip to do whatever you need—without running a software application.

This meeting of the minds means there are several opportunity areas for designers when it comes to an “All Programmable” environment for embedded vision: power efficiency from no OS, flexibility due to premade kits and a variety of applications enabled by a suite of cost-optimized devices.

No external OS, no problem
All Programmable solutions feature field programmable gate arrays, ICs that are configured post-manufacturing. This allows FPGAs to operate software applications out of a hardware implementation. In other words, a hardware system is dedicated to processing logic, so it doesn’t need an operating system

One of the biggest benefits of this for engineers is power efficiency. Because each processing path is parallel, operations aren’t all competing for the same processing resources. For embedded vision solutions, that means processing images in real time and turning those into actionable data that can stop a car on a dime or report safety data in an industrial factory—a big differentiator.

Multiple levels of hardware and software security also mean that privacy concerns are lowered while you are taking advantage of ARM® and FPGA architecture.

Hardware flexibility speeds time to market
All Programmable reference designs and hardware kits are bundled and ready off the shelf to customize to a specific engineering need. These bundles help speed your time to market.

No need to be an overnight expert in emerging technologies with deadlines looming. Peripherals like USB 2.0, Gigabit Ethernet and SD/SDIO can all be adapted on the processing system. Accelerators can be added to existing programmable logic allowing quick adaptations of kits to a lengthy list of applications. Despite integrating CPU, DSP, ASSP, and mixed signal functionality onto one device, dedicating hardware to processing logic reduces unnecessary power consumption, too.

While there’s less power needed, however, these bundles still deliver all the performance required depending on the cores, memory and application.

Embedded vision applications are nearly limitless
With a variety of cost-optimized devices with up to 866MHz cores and mid-range devices with up to 1GHz of core processing and from 1.8Mb to 26.5Mb of memory respectively, there’s a strong price-to-performance ratio. This means there’s a version of “All Programmable” that can work in a variety of high-end control applications.

  • Automotive: Autonomous driving depends on embedded vision to monitor drivers, the road and environment to eventually create safer roads.
  • Medical: Machine vision is crucial in treatment tools for hospitals, helping to create better diagnoses by doctors and more efficient scanners and x-ray machinery.
  • Industrial: Embedded vision powered by All Programmable SoCs can crop up in robotics, machine tools as well as conveyor, compressor, fan and pump controls across factory applications.

Still not convinced? Learn more about how embedded vision turns complexity to clarity.

3 ways All Programmable SoCs create opportunities | Avnet Silica

Display portlet menu


Developing multicamera vision systems for autonomous vehicles

One of the biggest challenges in developing vision systems for autonomous vehicles is handling the l...

Elements in the development of embedded vision systems for real-world applications

Improvements in processing performance, power consumption, image sensors, computer algorithms and ma...

How all programmable technology revolutionizes embedded vision

We’re driving toward new innovations in Embedded Vision – and the future of driving is only the begi...