NETRA-SOC Family Edge-AI Chips
NETRA-A2000
NETRA-A2000 is a Smart vision capable SOC with advanced real-time video analytics and vision processing capabilities. The chip features Netrasemi TEJ-NPU (Neural processor) core and a Vision processor and H26x video codec. It is powered by Cortex-A family multi-core processor with TrustZone security features for ensuring chain-of-trust for secure IoT applications. The chip features LPDDR4, PCIe3.0, USB3.1, eMMC, MIPI-CSI2 and many useful peripheral interfaces for communication, storage, sensing and control functions.
Target Applications:
- Video analytics
- Security & Surveillance
- Drones
- Robotics
- Industry 4.0
- Autonomous Vehicles
- Smart infrastructure
NETRA-R1000
NETRA-R1000 is a RISC-V based SOC with excellent power efficiency for smart sensor applications. The chip features Netrasemi TEJ-NPU (Neural processor) core and a rich set of peripheral IPs. It is powered by a 64Bit RISC-V processor with secure boot and other advanced security features. The chip features LPDDR, USB, eMMC and many useful peripheral interfaces for communication, storage, sensing and control functions.
Target Applications:
- Smart sensors
- Home/Building automation
- Smart Toys
- Industry 4.0
- Anomaly detection
- Retail & office
- Gesture
- Smart vision
NETRA-A4000
NETRA-A4000 is an excellent 64 TOPS ARM-based intelligent SOC with a chipset-based die-to-die integration technology. The chip features a Netrasemi TEJ-NPU (Neural processor) core and a rich set of peripheral IPs. It is specifically catered for high-end applications and will serve the edge-server market.
Target Applications:
- Autonomous Vehicles
- Robotics
- Edge Servers
- Smart City Platforms
- Anomaly detection
- Intelligent NVR
- ADAS
- Defense/ Aero-Space
NETRA SmartEdge Studio
AI Software Framework
NETRA-SOCs come with a powerful SDK, AI utilities, AI benchmark tools and a powerful Graph App builder that minimises the overall AI application development time using our chipsets. With the help of Graph builder tools one can enable complex heterogeneous acceleration engines to process the data in a graph stream pipeline with minimal or no coding effort. This not only helps to reduce the development time and faster market access, but also ensures best power-performance efficiency for the application. The Software framework supports AI development using models generated by standard libraries like Tensorflow, PyTorch, Caffe and ONNX. We also provide reference hardware kits and demo application suite for various target applications. Together with NETRA SmartEdge studio, SDKs, RDKs and demo suite, customers significantly reduce their time to market.