By modifying the decoder in CUDA SDK,
I had wrote a program to decode multiple video in the GPU.
And I can also get the correct result after decoding.
When I create more then 10 decoders,
the CUDA respone an error about creating fail.
It happened at two functions call :
hr=cuGLCtxCreate(&hContext, CU_CTX_BLOCKING_SYNC,hDevice); -> hr = 999
hr=cuvidCreateDecoder(&hDecoder,&info); -> hr = 300
I don’t know why this happended.
Cause I get different count of decoder in different computer with the same GPU.
Does anyone have idea of this issue ?