PCIe AI inference acceleration card MYS-ZU3EG-8E4D-EDGE

FPGAXilinx® Zynq® Ultrascale+™Arm® Cortex®-A53 Quad-core

The FZ3 Deep Learning Accelerator Card, leveraging the Xilinx Zynq UltraScale+ MPSoC XCZU3EG, incorporates a 4-core Cortex-A53 processor. Its measured performance can reach up to 1.2 TOPS, achieving 100FPS for MOBILENET under quantized pruning, exceeding CPU performance by 20 times, while consuming only 5-10W of power. When the model is not pruned or quantized, the performance of the accelerator card is still excellent. The FZ3 Card is a powerful deep learning accelerator card based on Xilinx Zynq UltraScale+ ZU3EG MPSoC which features a 1.2 GHz quad-core ARM Cortex-A53 64-bit application processor, a 600MHz dual-core real-time ARM Cortex-R5 processor, a Mali400 embedded GPU and rich FPGA fabric. Besides, it integrates 4GB DDR4, 8GB eMMC, 32MB QSPI Flash and 32KB EEPROM as well as many peripherals including USB 2.0, USB 3.0, Gigabit Ethernet, TF, DisplayPort (DP), PCIe interface, MIPI-CSI, BT1120 camera, USB-UART, JTAG, IO expansion interfaces, etc. The rich resources enable users to integrate intelligent hardware easily. The FZ3 Card is able to run PetaLinux 2020.1 and and provided complete BSP. It can support Xilinx Vitis Software development platform. It can also supports PaddlePaddle deep learning AI framework which is fully compatible to use Baidu Brain’s AI development tools like EasyDL, AI Studio and EasyEdge to enable developers and engineers to quickly leverage Baidu-proven technology or deploy self-defined models, enabling faster deployment. Typical applications are AI camera, AI computing device, robotics, intelligent car, intelligent electronic scale, patrol UAV and other embedded intelligent applications.