I am setting up my GPU server with 2 GPUs. When I got my server from manufacture, these GPUs were installed into the different CPUs’ sockets so there is no Peer-to-Peer access between 2 GPUs. I think P2P access has more advantages for accelerating my deep learning programs because of direct memory access between 2 GPUs. So, I am just wondering why my manufacture install my GPUs without P2P capability. Can anyone explain if there is any advantage of no P2P access between 2 GPUs?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Two GPU cards on a single machine | 3 | 1167 | September 29, 2016 | |
| How can I tell which NVIDIA GPUs will have P2P access to the same GPU on PCIe? | 6 | 8202 | January 20, 2025 | |
| Peer-to-Peer Access Fails between 2 GPUs | 3 | 5774 | July 7, 2017 | |
| Why cudaDeviceEnablePeerAccess is not default? | 1 | 1185 | December 11, 2019 | |
| Peer-to-Peer access GPUDirect | 0 | 6120 | August 8, 2011 | |
| Two way peer access between two GPU cards | 1 | 1000 | June 23, 2016 | |
| Multiple GPUs, Peer-to-Peer Question | 1 | 1251 | October 21, 2016 | |
| Peer-to-Peer Communication | 1 | 2458 | October 13, 2014 | |
| Failed to simpleP2P | 3 | 137 | August 26, 2024 | |
| simpleMultiGPU seems to work fine but simpleP2P fails | 0 | 1071 | December 14, 2017 |