I have went through the Nvidia series of GPU, i am trying to work on a encoder/transcoder using NVENC API.
- Where i can find the informations. All the spec in the GPUs has more Cuda and Tensor Cores, but only one Encoder and 2or 3 decode engines. Not sure What does that mean? Is it limited by 1 Encode engine or its the capacity of the cuda core should i consider.
- how may streams can a GPU model handle and how to calculate using the spec.
- where do i get the libraries and samples to try on my own.