The latest Azure HPC SKUs series, HB and HC, unleash the true power of Azure's performant and scalable HPC network interconnect by virtue of Single Root IO Virtualization (SR-IOV). These high performance VM series offer true InfiniBand capabilities and enable all InfiniBand based MPI libraries to run on Azure with full RDMA verbs.
To help HPC users seamlessly take advantage of all the platform features, the Azure HPC team offers a CentOS based HPC image in Azure Marketplace (published by Rogue Wave Software). This image, starting today with CentOS 7.6, includes optimizations and recommended configurations to deliver optimal performance, consistency, and reliability. This is the same image we recently used to set to run the public cloud's first ever 10,000+ core MPI job!
Below is the full list of configuration updates and software installations (on top of a vanilla Azure CentOS VM image):
The included configuration updates are based on optimization recommendations from vendors and partners, as well as learnings from common HPC workloads and usage practices in traditional HPC systems.
The CentOS HPC image also comes with a set of pre-installed software packages:
The MPI libraries are configured as environment modules. Users can select any preferred MPI flavor as follows:
module load mpi/<mpi-flavor>
Please refer to the corresponding CentOS-HPC kickstart file for a complete list of configuration updates and software configurations.
Azure HPC team aims to keep the CentOS HPC images up-to-date with latest software versions.
Deploying the CentOS HPC Image
The CentOS HPC images are available from Azure Marketplace, and it can be deployed through a variety of deployment vehicles (CycleCloud, Batch, ARM templates, etc).
Go ahead and give this image a try for your HPC workloads! We're all ears for your feedback, including additional recommendations for optimizations or included software, even other OSes Azure should optimize for HPC purposes.
Note: This image is intended *only* for the SR-IOV enabled Azure instances (such as Azure HB, HC, etc.).
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.