- iWave Systems’ evaluation kit (EVK) incorporates tools and libraries for reduced complexity in developing AI solutions
- Arm-based FPGA architecture provides advantage of using the Arm cores as a regular SoC for high-level management functionalities
As AI is increasingly becoming part of various application areas, there is an increasing requirement for adaptive intelligent solutions capable of AI workloads, particularly in edge devices. FPGA based development solutions offer exceptional inference capability with the help of complex neural networks and deep learning algorithms while maintaining flexibility, low latency and power efficiency.
Now, iWave Systems has announced the successful implementation of the open-source Xilinx Vitis™ AI stack on iWave Zynq® UltraScale+™ MPSoC evaluation platform. Vitis™ AI consists of tools, libraries, pre-optimized models and example designs, which reduces the complexity and overall effort for the development of AI design solutions. The evaluation platform with multiple connectivity, camera and support for Vitis™ AI offers a complete and well-integrated solution enabling the development of various Machine Learning and AI applications.
The powerful Zynq® UltraScale+™ MPSoC EVK offers a complete and well-integrated solution to jumpstart FPGA-based AI applications. The Vitis™ AI Stack includes advanced pre-optimized deep learning models from the mainstream frameworks such as Tensor-flow, Caffe, Darknet and Computer Vision Libraries. Vitis™ AI Stack supports C++/python APIs, allowing faster development of AI applications even without in-depth knowledge of FPGA and deep learning.
The solution incorporates a configurable deep-learning processor unit in the PL of the Zynq MPSoC chip. The DPU is an AI inference engine dedicated to Convolution Neural Networks such as VGG, SSD, Yolov2/v3, FPN, Resnet50 and others. The DPU AI inference engine provides multi-dimensional parallel architecture capable of performing major convolutional calculations through deep pipelined computing engines.
The EVK supports interfaces to connect 8 IP cameras, multiple USB cameras and SDI Camera, providing the ability to capture multi-angle high-resolution video/image frames. The solution also supports high-speed connectivity options such as Dual Gigabit Ethernet and 3G/4G/5G via an M.2 expansion slot. There is also a provision for extended storage for an mSATA / NVMe SSD via the M.2 expansion slot. Additional interfaces such as FMC, SATA, SDI Video In & Out, USB3.0, SFP+ and a host of other interfaces allow for a wide range of networking options helping customers towards various use cases based on the on-premise architecture and requirements.
The EVK is built around a highly adaptive Zynq® UltraScale+™ MPSoC that features a heterogeneous Arm® + FPGA architecture. This provides users with the advantage of using the Arm cores as a regular SoC for high-level management functionalities, while the freely programmable FPGA can be configured to dynamically accelerate functions for real-time AI inference.
The Zynq® UltraScale+™ MPSoC has an integrated Arm Mali-400MP2 GPU engine, H.264/H.265 VCU and other core signal processing, networking and transceiver sub-systems that further enhances the AI deployment efficiency. The MPSoC also incorporates multiple levels of security, increased safety and advanced power management, enabling an optimized solution for AI design development.
Through the support of key features such as the Zynq® UltraScale+™ MPSoC, connectivity, camera and the support for the Xilinx Vitis™ AI Stack, the EVK qualifies as a perfect fit for the development of various machine learning and AI applications such as object detection, image classification, face detection/recognition, ANPR, video surveillance and many more.