Multi-Instance GPU
NVIDIA Multi-Instance GPU User Guide r570 documentation
Compute Instance
Compute Instance可以几个lanes的共享一块SM
MPS带宽有隔离,MIG带宽有隔离
MIG用不了PCIe和NVLink
CUDA MPS is supported on top of MIG. The only limitation is that the maximum number of clients (48) is lowered proportionally to the Compute Instance size.