Efinix is a company that makes Field Programmable Gate Arrays (FPGAs). It was started 10 years ago by a team of veterans of the FPGA industry who believed that they had a much more efficient architecture for FPGA devices. That belief has been born out and today Efinix makes the world’s most efficient FPGA devices resulting in cost effective, low power and small footprint devices that are ideally suited for space and power constrained applications such as edge compute or machine vision.
Interview with Mark Oliver, VP Marketing at Efinix.
Easy Engineering: What are the main areas of activity of the company?
Mark Oliver: Our FPGA devices deliver a blank canvas on which to design custom silicon applications. They are used in every conceivable application but, due to the included MIPI interfaces, find particular use in machine vision and camera applications. They bring an efficient compute platform to these applications where their small size and low power consumption results in a very compact end design. The customizable nature of the FPGA along with its efficiency and density means that ample room is left for the designer to differentiate the product with custom image processing algorithms or artificial intelligence.
E.E: What’s the news about new products?
M.O: We have recently announced our latest member of the Titanium family of products. The Ti180 has 180,000 logic elements and provides unprecedented power and performance points. It has a hard LPDDR4/4X interface and 2.5G MIPI. It is compatible with the Efinix RISC-V compute core and the recently announced TinyML platform that speeds time to value for designers innovating in the space of artificial intelligence
E.E: What are the ranges of products?
M.O: At Efinix we have two families of FPGA product. Our value range is called Trion and is fabricated on a 40nm process node. The performance range is called Titanium and is fabricated on a 16nm node. Together with the design tools, libraries and downloadable IP, they provide an end-to-end solution from the most space and power constrained applications up to the most compute intensive embedded systems.
E.E: At what stage is the market where you are currently active?
M.O: The world is seeing a gradual change in the semiconductor industry. Moore’s law that has predicted the increase in compute performance with reducing prices, is slowing down. Increasingly it is becoming more cost prohibitive to create custom silicon solutions. Efinix FPGAs are the ideal solution providing custom silicon type designs with zero risk and NRE in a platform that is priced for high volume manufacture.
E.E: What can you tell us about market trends?
M.O: There is a trend in the industry to decentralize compute away from the data center towards the edge in order to achieve lower latency and better quality of service. This is seen particularly with the rollout of 5G where a new breed of applications is being designed to capitalize on the high bandwidth and low latency. The result is an increased demand for heterogeneous compute capability in small, low power footprints that can be deployed as close as possible to edge devices if not actually inside those devices. Efinix FPGAs are ideal for implementing heterogeneous compute architectures at the edge where custom hardware can be instantiated to process edge data in real time. The Efinix RISC-V core with custom instruction acceleration provides all the compute power needed for implementing AI right out at the far edge.
E.E: What are the most innovative products marketed?
M.O: The Efinix titanium line is the most innovative FPGA on the market right now. With a totally re-imagined architecture it delivers highly efficient designs in a small, low power footprint. Coupled with the software configurable RISC-V core and custom instruction acceleration, Titanium is the ideal architecture for space and power constrained applications.
E.E: What estimations do you have for the rest of 2022?
M.O: Efinix will continue to innovate going forward delivering unprecedented value in our two FPGA product lines. Through SoC templates and acceleration “sockets” we will continue to revolutionize the way designers access this groundbreaking technology.