Interacting with Other Engines

The cuPVA runtime facilitates powerful interoperability with other compute engines, allowing for the construction of complex, multi-stage image and data processing pipelines.

This section explores how cuPVA can be integrated with CUDA for GPU-accelerated tasks and with NvSci for standardized data and synchronization across different NVIDIA hardware blocks and APIs. These interoperability features enable efficient resource utilization and streamlined workflows, especially in safety-critical systems.

This section includes the following tutorials:

• CUDA Interoperability

• NvSci Interoperability