Azure

This guide will walk through how to set up a single Azure Instance with multiple GPUs. This guide is meant for testing purposes only.

Overview

Before starting, we recommend having the following items ready:

• Azure account

• Azure subscription and quota capable of starting GPU instances such as NC12 or better

• SSH public key

• SSH/SCP programs on local computer (eg., Putty/Winscp on Windows)

• M-Star license file

Basic tasks that need to be completed:

• Start up a VM with GPU hardware

• Install CUDA

• Install Open MPI

• Install M-Star CFD

• Run test case

Additional resources:

• Azure documentation

• CUDA documentation

• M-Star documentation

Selecting an Azure instance type

Last updated Dec, 2023

For users that need to run the entire workflow on the remote resource (remote desktop enabled), including Pre, Solve, and Post, this will require a remote desktop that has OpenGL 4+ support as well as a higher CUDA compute ability. Therefore we recommend using one of the following instance types: NV-Series, NVv3-Series, NVadsA10 v5-series, NCasT4_v3-series. For the best performance, the NVadsA10 v5 series should be the first choice.

For users that only need to run the solver from the command line (no remote desktop), there are more options. In this scenario the computing resource does not require high OpenGL display support. Instances that are compatible are: NCv3-series, ND A100 v4-series, NDm A100 v4-series, NDv2-series, ND H100 v5-series.

More resources:

• Configure GPU acceleration for Azure Virtual Desktop

• Install NVIDIA GPU drivers on N-series VMs running Windows

• Accelerate graphics-heavy workloads using NVads A10 v5 Azure

• GPU optimized virtual machine sizes

Create a virtual machine

1. Go to the Azure portal to start the creation of a new virtual machine and fill out the form as follows (some options may be different for you):

   • Subscription: Pay-As-You-Go (your choice may differ)

   • Resource group: Testing (your choice may differ)

   • Virtual machine name: testing

   • Image: Ubuntu 18.04

   • Size: Standard NC12 (Select a multi-GPU node from the series NC, NCv2, or NCv3)

   • Authentication Type: SSH public key

   • Username: test

   • SSH public key: Paste your SSH public key here

   • Public inbound ports: Allow selected ports

   • Select inbound ports: SSH

2. Click Review and create.

3. Click Create SSH to the new virtual machine.

Install M-Star CFD

See Linux (Single Node).

Troubleshooting Pre and Post startup

See Troubleshooting M-Star Pre/Post start up for additional information