- Posts: 1
- Joined: Wed May 22, 2019 12:50 am
- Real Name: minghai lu
Wed May 22, 2019 1:26 am
hello:
we read the GPUdirect sample of decklink.However it seems that it should create a shared pinned memory, so the data transfer is that decklink card -> shared pinned memory -> video memory.
Actually it still consume time for us,and we want the transferred data like that decklink card -> video memory.
There is an GPUdirect RDMA technology or GPUdirect P2P of NVidia card.So we want to know whether decklink cart support GPUdirect RDMA ,if so,how to realize it in the decklink card. I beg give us some tutorials.Thank you.
Or if there exits other methods about how the decklink card transfer frames directly into the video memory of nvidia ,please give us some tips ,thank you.
we read the GPUdirect sample of decklink.However it seems that it should create a shared pinned memory, so the data transfer is that decklink card -> shared pinned memory -> video memory.
Actually it still consume time for us,and we want the transferred data like that decklink card -> video memory.
There is an GPUdirect RDMA technology or GPUdirect P2P of NVidia card.So we want to know whether decklink cart support GPUdirect RDMA ,if so,how to realize it in the decklink card. I beg give us some tutorials.Thank you.
Or if there exits other methods about how the decklink card transfer frames directly into the video memory of nvidia ,please give us some tips ,thank you.