I’m currently trying to decode/render multiple videos using the NVIDIA decoder and D3D11 and I’m finding that on one machine the decoder is producing garbage and the decoding status is constantly returning errors. Here’s my setup:
*Multiple windows that each have their own swap chain
*1 master render thread that is responsible for creating/using the swap chain
*Multiple decoders that share a single context (using a floating context & a lock that is associated with the context which I’m using whenever I call any cuda functions)
I’m finding that as soon as I try and use more than 2 swap chains, the decoders start returning errors (cuvidGetDecodeStatus == cuvidDecodeStatus_Error_Concealed).
I’m only seeing this issue on one machine so far. Any ideas?