As Artificial Intelligence moves to the edge requiring algorithms be processed locally on a hardware device. Computation at the edge requires a device that can handle a sufficient workload, while maintaining low power because of thermal constraints. Microchip Mid-range FPGAs deliver high throughput at a low power allowing them to be a great option for implementing Artificial Intelligence and Machine Learning. With the introduction of the VectorBlox Accelerator Software Development Kit the issue of enabling software engineers to be able to scale neural networks with FPGA-based AI solutions becomes solved. The PolarFire SoC Icicle Kit with it is RISC-V hardened Processer will address the need for processing to be real-time execution in AI/ML at the edge. The Hello FPGA kit is a low-cost board developed to help those with low-mid knowledge on FPGAs begin their journey, it is packed with demos in Image Processing, Signal Processing, and Artificial Intelligence.
1. Vectorblox
2. Hello FPGA Kit
3. PolarFire Video Kit
4. FPGA Security
1. VectorBlox
Buliding out Microchip's FPGA AI/ML portfolio the new CoreVectorBlox Neural Network IP and VectorBlox Accelerator Software Developement Kit makes designing low power and small form factor AI/ML applications easier. The CoreVectorBlox IP is a flexible neural netork accelerator that uses a overlay approach, it has the ability to run different networks on a single instantiation without needing to be resynthesized. The VectorBlox Accelerator SDK has different tools that can compile trained neural networks with TensorFlow and ONNX into a Binary Large Object.
2. Hello FPGA Kit
Buliding out Microchip's FPGA AI/ML portfolio the new CoreVectorBlox Neural Network IP and VectorBlox Accelerator Software Developement Kit makes designing low power and small form factor AI/ML applications easier. The CoreVectorBlox IP is a flexible neural netork accelerator that uses a overlay approach, it has the ability to run different networks on a single instantiation without needing to be resynthesized. The VectorBlox Accelerator SDK has different tools that can compile trained neural networks with TensorFlow and ONNX into a Binary Large Object.
Downloads
Watch Video of Hello FPGA Kit Demonstration
3. PolarFire® Video Kit
PolarFire® Video Kit uses it's numerous display interfaces and dual camera sensors to open up new possibilities to implement systems that rely on visual data to make decisions across a broad spectrum of applications. Drones, Machine Vision, Thermal Imaging, Gaming, and Video Surveillance are all applications where having the ability to evaluate 4K image processing while maintaining low power can come in handy. The kits interfaces eg. MIPI CSI-2 TX, MIPI CSI-2RX, HDMI 1.4Tx help to make prototyping effortless in imaging and video spaces.
Downloads
4 Imaging and Video Solutions for Smart Embedded Vision Brochure
4. FPGA Security
Microchip FPGA takes Cyber security as a #1 concern for connected devices on the edge. With features that prevent overbuilding and cloning, have tamper detection and responses, while making sure the hardware stays secure with a Root of trust, it is these combined security features that make Microchip PolarFire FPGA one of the most advanced secure programmable FPGA.