[Feature]: Support H100 in OCI

[jvstme]

Problem

Oracle Cloud provides on-demand bare metal 8xH100 instances ( BM.GPU.H100.8). Currently, these instances cannot be provisioned with dstack.

Solution

Support the BM.GPU.H100.8 shape.

Known challenges

1. [Bug]: fabricmanager-dependent GPUs do not work in AWS, GCP, Azure, OCI #2348
2. At the time of testing, the OCI API was reporting that BM.GPU.H100.8 is out of capacity.

    #  BACKEND  REGION          INSTANCE       RESOURCES                                      SPOT  PRICE                
    1  oci      eu-frankfurt-1  BM.GPU.H100.8  256xCPU, 2048GB, 8xH100 (80GB), 50.0GB (disk)  no    $80    not available
   

   But in fact it was possible to start BM.GPU.H100.8 via both the web console and the API. If this will be reproduced later, we'll need to research how accurate the OCI capacity reports are and reconsider using them.

Workaround

No response

Would you like to help us implement this feature by sending a PR?

Yes