No matter what I do, nppGetStreamContext(&m_nppStreamContext) results in a seg fault. (I have NppStreamContext m_nppStreamContext; declared in the header.)
Other related calls work fine. i.e. cudaSetDevice(), cudaStreamCreate(), nppGetGpuNumSMs(), nppSetStream(), nppGetStream(). I call all of them before calling nppGetStreamContext().
I can get a cudaStream_t without issue, but not a NppStreamContext. I tried manually building it as well, but that also results in a seg fault. i.e.:
m_nppStreamContext = {m_cudaStream, deviceId, properties.multiProcessorCount, properties.maxThreadsPerMultiProcessor, properties.maxThreadsPerBlock, properties.sharedMemPerBlock, properties.major, properties.minor, 0, 0};
How can I troubleshoot this issue further?