Digi-Key at embedded world 2023 with Efinix

Formed in 2012, Efinix have built a team that are “changing the dynamics of how the FPGAs work, and their functionality, and the applications that they go into.”

Their Quantum architecture is designed to be the leading programmable technology, delivering the best power, performance, and area for silicon products, with the goal to provide the most flexible programmable technology, tailored for compute-intensive applications including compute acceleration, machine learning, and deep learning.

What makes Efinix FPGAs different?

 “The standard of the FPGA architecture has not changed over the last 20 years. There is fixed routing and fixed logic and you have to provide a lot of routing resources to connect from A to B.

“Seeing all the disadvantages and the bottlenecks out of this architecture … our new XLR [eXchangeable Logic and Routing] technology combines both together. We have an XLR cell which gives you the ability to do logic implementation … and at the same time you can integrate routing resources into the same cell. So, the cell can do logic, routing or both together.”

The industrial and automotive market are two areas of industry that Girardi believes are seeing the most customer interest in FPGAs.

How will the embedded space evolve?

“The embedded space is evolving tremendously, especially these last few years.

“With … Efinix FPGAs you can get onto your design right away, you can change it, you can debug it and you can bring it into the market … because of our architecture we’ve reduced that gap to A6”.

To programme an FGPA, you need to know the language. Efinix are providing a solution where the FPGA can be designed with C-software developers. Efinix RISC-V solution FPGA enables support of acceleration of the design and a custom instruction set of a RISC-V.

Efinix FGPAs are ideally suited for Edge applications

“You want a robust solution, but you always want extremely low power so the [Efinix] FPGA is ideally suited to Edge applications. You can offload some of the processing power but there is very low latency and high throughput of performance all the way down the end user applications.”

Efinix have also released a TinyML Platform for artificial intelligence (AI) acceleration on its range of FPGAs.

The TinyML Platform works with the high performance and Linux-capable Sapphire RISC-V processor. Comprising a model profiler and graphical user interface for accelerator selection, the TinyML Platform delivers an accelerated AI capability to the quad-core-capable Sapphire core.

“[Our] TinyML, or artificial intelligence, is based on the RISC-V processor. We use the functionality of the RISC-V processor which means we are using custom instruction and acceleration for the different function in an artificial intelligence network.”

“We’re trying to make it as easy as possible for design engineers to implement whatever function they’re trying to do on the FPGA.”