Data streaming

REDsdk can be used in applications that render large assemblies and that require data streaming. First, the preferred rendering architecture for these applications can be the Multi-threaded performance critical applications: one thread will be in charge of feeding REDsdk with new streamed data, while the rendering takes place in parallel, on the rendering thread.

The application must take care of image management, as detailed here: Modifying images during a draw and there: Image management, as REDsdk is synchronous on images.

Then, data streaming may imply on-the-fly operation on existing datasets. Material batching may be key to reach a good performance level in performing selective edition of GPU memory batches.