GreenWaves Intros Open-Source AI Processor GAP8


GreenWaves Technologies has unveiled a new open-source processor in the form of the GAP8 which comes with a big focus on AI. The company, which was founded back in 2014, is based in Grenoble, France, and currently focuses on the development of low-energy solutions for AI devices. More recently, the company raised €3.1 million to turn their work into a commercial product, with the result being the new GAP8 IoT application processor and the availability of the GAP8 software development kit.

The company’s new processor is based on the RISC-V open-source processor architecture, with the focus being on handling low-power AI processing in sensory devices that other mainstream chips would not typically be designed to handle specifically. In fact, GreenWaves has designed the processor with image, sound and vibration analysis at its heart, with a number of new algorithms being included in order to execute a wide variety of tasks. These tasks will also consume minimal amounts of energy due to the integrated 8-core cluster that is coupled with a separate core designed to handle any pre-analysis communication, control, and information. It is because of this low power consumption that GreenWaves has designed the processor with battery-powered devices in mind, although it hopes the chip will result in a number of new connected products with support for artificial intelligence such as smart toys, certain wearables, or even the implementation of always-on facial recognition in mobile devices. However, the new processor isn’t just energy-efficient, but also relatively affordable, with the handling of machine vision potentially costing less than $15 to implement. The product should also help relieve pressure on networks due to the fact that all processes will happen wherever the sensors are placed, removing the need for a secondary product while also reducing the costs of data management and speeding up the processing, according to the company.

GreenWaves’ new processor can operate on as little as few milliwatts, while other chips on the market typically consumer hundreds of milliwatts. This coupled with the reduced costs that come with its open-source nature could easily give a number of startups the potential to build their products on much tighter budgets, while mainstream companies could also opt to use the chip in the hope of saving huge amounts of money that would usually be spent on purchasing intellectual property for their own chips. The GAP8 SDK is already available, while pre-orders for GAP8 evaluation boards have also commenced.