Skip to main content

Overview

OGC Virtual Machines provide a powerful and flexible way to run AI/ML workloads.

  • GPU Availability - Through OGC you can get the widest variety of GPU types (e.g. V100, V100s, A100, L4, L40s, H100, H100 SXM and more!) and locations across the globe.
  • Fractional GPUs - VMs come in many configurations of GPUs, from small fractions of a GPU (e.g. 1/20th) up to multiple GPUs in a single node.
  • Per Minute Billing - Billing is usage-based and charged per minute.
  • Advanced Features - Ability to Suspend, Resume and Restart machines to better manage usage and optimise spend.

Use Cases​

  • Ideal for large ML inference/training workloads looking to leverage the most powerful GPUs (e.g. 8x H100 SXM).
  • Suitable for customers looking for flexible pricing or fix term contracts.
  • Perfect for running PoC and experiments due to on-demand pricing and fractional GPUs.

Resources​

  • Run Ori's Benchmarking framework BefOri.
  • Read about how Billing works.
  • Review your available Quotas.

Contents