NVIDIA IndeX is a platform for writing real-time visualization solutions giving unique insights into large and complex scientific data. The NVIDIA IndeX platform has single and multi GPU support built-in and facilitates easy use of large GPU cluster in data centers (on-premise), in the cloud or on Supercomputers for scalable data visualization. The NVIDIA IndeX platform comes with a variety of volume dataset types such as dense or sparse, regular or unstructured volume dataset types serving different domains such as seismic, microscopy, astronomy and cosmology. Its versatility comes from the wealth of interfaces, which make it easy to
- Produce high-fidelity custom and domain specific visualization through the CUDA-based programmable XAC API,
- Process and analyse the visualization data in parallel at a high performance through tasks that a user can implement and that are schedules to the potentially distributed many GPUs automatically,
- Invoke deep-learning inference operations that may rely on the user's deep-learning framework of choice, and
- Trigger parallel data imports of virtually any volume dataset format from any data location.
The NVIDIA IndeX platform eases the integration into todays visualization solutions such as Kitware's Paraview or the development of new tailor-made visualization applications, the implementation of visualization servers or cloud services and visualization apps that are accessible around the globe. The core NVIDIA IndeX library, the SDK, comes with a sophisticated C++ API. An NVIDIA IndeX application layer on top of the SDK extends the core functionalities, builds the NVIDIA IndeX platform environment, includes a variety of components such as a video streaming service and canvas and command management, and exposes a mechanism for custom plugins either implemented by NVIDIA and shipped as part the NVIDIA IndeX package or implemented by users, the visualization community and customers for their specific use of NVIDIA IndeX. The NVIDIA IndeX platform has a C++ API and includes a WebSocket API for easy integration into web applications (e.g., HTML viewers in browsers) or easy use of a visualization service in Jupyter notebooks.
The NVIDIA IndeX software distribution consists of three parts:
- The NVIDIA IndeX SDK library and C++ API,
- The NVIDIA IndeX application layer libraries, the C++ APIs, the application layer source code and the WebSocket API,
- The NVIDIA IndeX viewer, which is based on the NVIDIA IndeX platform, shall support users to get started with their own data and is commonly used in diverse demo applications and prove-of-concept studies.
- This documentation set that includes references for the APIs used by NVIDIA IndeX as well as descriptions of the software source code examples.
See the section Getting started for information on licensing, system requirements and using an example viewing application to test your IndeX installation.