Intel® FPGA AI for Edge White Paper

Intel FPGA AI Suite enables FPGA designers, ML engineers, and software developers to efficiently optimize AI-enabled designs based on Intel FPGAs. Utilities in the Intel FPGA AI Suite speed up FPGA development for AI inference using familiar and popular industry frameworks such as TensorFlow or PyTorch and OpenVINO toolkit, while also leveraging robust and proven FPGA development flows with Intel® Quartus® Prime software. Three Intel FPGA and SoC families are especially well adapted to implementing AI workloads. These three families are the Intel® Stratix® 10 NX FPGA and the newest members of the Intel® Agilex™ FPGA family: the Intel Agilex D-series FPGAs and the new Intel Agilex family code named “Sundance Mesa.” These Intel FPGA and SoC families incorporate specialized DSP blocks that are optimized for tensor math, which is fundamental to performing AI calculations more quickly.