As a sanity check, does this happen when you use the device names as MIG- like described by https://docs.nvidia.com/datacenter/tesla/mig-user-guide/#running-containers ?
ScottE
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Training in TAO 3.0 in DGX inside docker-compose for multi-gpus MIG | 5 | 1442 | March 1, 2022 | |
| Docker container unable to see more than 1 GPU | 0 | 463 | February 13, 2020 | |
| Getting the Most Out of the NVIDIA A100 GPU with Multi-Instance GPU | 11 | 1641 | January 19, 2023 | |
| Is it possible to have one MIG device recognized as one GPU in a docker container? | 0 | 57 | May 16, 2025 | |
| Any setting(enivronment variable etc) to not all GPU's in Multi-GPU hardware? | 2 | 609 | June 29, 2011 | |
| Pytorch issue with MIG instances created on A100 | 1 | 1146 | June 30, 2022 | |
| MPI with Multi-Instance GPU (MIG)? | 1 | 1008 | July 15, 2020 | |
| Nvidia DIGITS Container Image | 4 | 611 | April 26, 2018 | |
| Running code on iGPU vs dGPU | 9 | 2039 | October 12, 2021 | |
| Dividing NVIDIA A30 GPUs and Conquering Multiple Workloads | 0 | 372 | August 30, 2022 |