Introduction to Holoscan SDK
🚧 Documentation is under development
The Holoscan Platform guide is currently under active development. Some sections may be incomplete or change without notice.
Questions? Contact RidgeRun or email to support@ridgerun.com.
The NVIDIA Holoscan SDK is a software development kit designed to enable real-time AI processing for streaming data applications, especially in the fields of healthcare, robotics, and industrial systems. It is part of the broader NVIDIA Holoscan platform, which provides a complete edge AI computing framework by integrating hardware (like Jetson and IGX Orin platforms), software, and APIs.
The Holoscan SDK has some interesting features, such as:
- Graph-based Processing: Enables developers to build and manage data processing pipelines using a directed acyclic graph (DAG) model.
- Streaming Data Support: Optimized for real-time sensor and video stream data, allowing continuous processing and low-latency responses.
- Multi-domain Integration: Supports various data types and operations, including image processing, sensor fusion, and AI inference.
- GXF Runtime Engine: Uses the GXF (Graph Execution Framework) for modular and scalable graph execution.
- AI Inference Integration: Compatible with NVIDIA AI tools like TensorRT, DeepStream, and CUDA, facilitating high-performance inference workloads.
The Holoscan SDK is an integral framework that combines multiple NVIDIA libraries, such as cuPY, RAPIDS, TensorRT, and others:
Holoscan SDK accelerates the development of AI-powered applications that need to process complex sensor data in real time, particularly at the edge.