As long-term supporters of using open source to build out developer communities and generate new business opportunities, we were very excited to work with QuickLogic on enabling the first fully open source FPGA platform - the EOS S3-based QuickFeather.
Most noteworthy is QL’s leadership in embracing open source FPGA tools; this is not the only but the most important reason why it’s so exciting to see and support this Crowd Supply campaign.
We believe FPGAs are highly underrepresented in the world’s processing platform landscape, but their true potential can only be unleashed using open source… everything. IP, operating systems (where CPUs are involved, which is everywhere these days), simulation, hardware designs and, of course, FPGA tools. QuickLogic understood that, and worked with us - and now also the open source community - to create an officially open source platform which has nothing to hide, first time in FPGA history.
Thus, buying a QuickFeather in this Crowd Supply campaign is not just buying a developer platform - it’s getting a piece of computer history.
But it’s an extremely interesting dev kit, too. Machine Learning tasks performed on low-power devices like the EOS S3 found in QuickFeather, dubbed TinyML, is a dynamically growing area where we believe FPGAs will play an important role thanks to their ability to be configured to fit a specific case exactly.
A tiny Cortex-M4 FPGA SoC coupled with all the right sensors, enabling you to build a flexible FPGA-based ML accelerator, is a perfect development platform. And opening up the TinyML space to tinkering with AI acceleration by software developers thanks to open source FPGA tools is exactly what pushes it from "very interesting" to "awesome!"
Thus, we are excited to see this project gaining traction on Crowd Supply, and all the more proud for our role in developing the Zephyr support, basic FPGA IPs, the FPGA tools, and hardware.
If you’re on the fence whether QuickLogic’s QuickFeather is worth taking a look, you can also experience it in Renode - our open source simulation framework, whose version 1.10 (just released!) includes support for more EOS S3 peripherals, including DC, SPI DMA, Packet FIFO, and FFE, so that you can develop code, ML algorithms, and Continuous Integration for the platform even more quickly.
But of course, to put it in your next project or product, you need a physical board. We encourage you to go and get one, and experience the first vendor-supported and truly open source FPGA board with an ecosystem.