Score:0

Enablement of MIG feature in A100 attached to a VM

vn flag

I have a VM with single A100 attached. VM created with GCP 'çreate instance' method (Singapore region). The instance was not created with gpu-partition-size=1g.5gb

  • NVidia driver v470.57.02
  • OS RHEL7.9
  • CUDA v11.4.1.

The driver is working because nvidia-smi command gives expected output showing a single GPU ID 0 with 40G memory. I want to now enable MIG feature but executing the command gives error 'Not Supported'. Rebooting and reloading does not help. I am following the nvidia guide for A100 https://docs.nvidia.com/datacenter/tesla/mig-user-guide/.

Can A100 be configured for MIG once a VM has been instantiated without MIG partitioning? Is it that to use MIG feature, a VM need to be created using --accelerator type=nvidia-tesla-a100,count=1,gpu-partition-size=1g.5gb only?

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.