Azure offers several virtual machine series and sizes that are designed and optimized for compute-intensive tasks. They are also known as compute-intensive instances. At the time of writing this book, Azure offers the A8-A11, the N-series, and the H-series, which all support HPC workloads.

These series and sizes consist of hardware that is designed and optimized for compute-intensive, graphics-intensive, and network-intensive applications. They are best suited for modeling, simulations, and HPC cluster applications:

• The A-series offers RDMA networking, which provides ultra-low-latency and high bandwidth networking.
• The N-series is aimed at graphics-intensive and compute-intensive applications. They consist of different NVIDIA Tesla GPUs that are well suited for deep learning applications, gaming applications, or virtualization.
• The...

Microsoft provides an HPC Pack for Windows Server 2012, 2016, and Linux machines. This is a free offering and you can use this to create HPC clusters on your on-premises servers and Azure VMs.

You can install HPC Pack on a Windows or Linux Server. These machines will automatically become the head node of the cluster. You can then add additional nodes to the cluster and run a job on it. This job will be distributed across all the available nodes automatically.

It offers the following additional features:

• Hybrid Cluster: You can set up hybrid clusters using on-premises servers and Azure VMs
• HPC Cluster Manager: A tool for managing, deploying, and configuring HPC clusters
• PowerShell: You can use HPC PowerShell to manage, configure, deploy, add, and execute jobs on the cluster

You can use HPC Pack for designing effective cloud-native HPC solutions and hybrid...

Azure Batch is a service that helps developers scale their workloads over virtual machines (Windows or Linux) or containers, without the need to manage the infrastructure. With Azure Batch, you can run large-scale, parallel and HPC applications efficiently in the cloud. Batch computing is most commonly used for applications that regularly process, transform, or analyze large volumes of data. Typically, HPC applications that run on Azure Batch include deep learning applications, image rendering applications, media encoding applications, and Monte Carlo simulations. Azure Media Services uses Azure Batch internally for media encoding as well.

The Azure Batch service uses Azure compute as its infrastructure, which means you can use both Windows or Linux VMs to host your applications or workloads. Azure uses the A8/A9 series VMs with RDMA networking internally for the nodes...

In this chapter, we have covered the compute-intensive applications objective. We've covered how to design HPC and other compute-intensive applications using Azure services, how to determine when to use Azure Batch, and how to design stateless components to accommodate scale and Containers on Azure Batch.

In the next chapter, we will cover the web application objective.

Answer the following questions to test your knowledge of the information in this chapter. You can find the answers in the Assessments section at the end of this book:

1. You want to create HPC clusters in your on-premises environment. Should you use Azure Batch?
   1. Yes
   2. No

2. For creating high-performance compute VMs, should we use the H-series?
   1. Yes
   2. No
3. To automate the creation of Azure Batch processes, should we use ARM templates?
   1. Yes
   2. No

You can check the following links for more information about the topics that are covered in this chapter:

• High performance compute VM sizes: https://docs.microsoft.com/en-us/azure/virtual-machines/windows/sizes-hpc?toc=%2Fazure%2Fvirtual-machines%2Fwindows%2Ftoc.json
• Options with HPC Pack to create and manage a cluster for Windows HPC workloads in Azure: https://docs.microsoft.com/en-us/azure/virtual-machines/windows/hpcpack-cluster-options
• Batch Documentation: https://docs.microsoft.com/en-us/azure/batch/
• HPC, Batch, and Big Compute solutions using Azure VMs: https://docs.microsoft.com/en-us/azure/virtual-machines/linux/high-performance-computing
• Deploy applications to compute nodes with Batch application packages: https://docs.microsoft.com/en-us/azure/batch/batch-application-packages
• Solution architecture: On-premises HPC implementation bursting to Azure: https...