NVIDIA Clara Holoscan

Introduction

• Overview
• Relevant Technologies
  
  • Rivermax and GPUDirect RDMA
  • Graph Execution Framework
  • TensorRT Optimized Inference
  • Interoperability between CUDA and rendering frameworks
  • Accelerated Image Transformations

Setup

• SDK Installation
• Additional Setup
• Third Party Hardware Setup

Using the SDK

• Overview
• Holoscan Core Concepts
• Holoscan by Example
• Creating an Application
• Creating Operators
• Graph Execution Framework (GXF)

Applications and Operators

• Holoscan Sample Applications
  
  • Endoscopy Tool Tracking Application
  • Hi-Speed Endoscopy Application
  • Ultrasound Segmentation Application & Customization
  • Multi AI Application
  • Bring Your Own Model (BYOM)
• Holoscan GXF Extensions
  
  • GXF Built-in Extensions
  • Holoscan SDK GXF Extensions

Modules

• Logging
  
  • Defining the log level
  • Calling the logger
• Visualization Module
  
  • Overview
  • Concepts
  • Usage
  • Layers
  • Using a display in exclusive mode
  • Cuda streams
  • Reading the framebuffer
• Inference Module
  
  • Overview
  • Parameters and related Features
  • Usage

API

• Holoscan C++ API
  
  • Namespaces
  • Macros
  • Classes
  • Enums
  • Functions
  • Typedefs
  • Variables
• Holoscan Python API
  
  • holoscan.conditions
  • holoscan.config
  • holoscan.core
  • holoscan.executors
  • holoscan.graphs
  • holoscan.gxf
  • holoscan.logger
  • holoscan.operators
  • holoscan.resources
• Holoviz API
  
  • Enums
  • Functions

Performance Tools

• Video Pipeline Latency Tool
  
  • Requirements
  • Installation
  • Example Configurations
  • Operation Overview
  • Graphing Results
  • Producers
  • Consumers
  • Troubleshooting

Links

• User Guide PDF
• NGC Containers
• Github Repository
• Previous releases