Azure HPC documentation

Official Azure high-performance computing (HPC) documentation. Learn about Azure HPC's various workload orchestration services that integrate with compute, network, and storage resources.

• High-Performance Computing on Azure
• Migrate HPC workloads to Azure
  • Get started
    • Plan your on-premises HPC migration to Azure
  • Migration guide
    • Build a proof-of-concept HPC cluster on Azure
    • Production deployment guide
      • Overview
      • Set up resource groups, networking, and foundational storage
        • Resource group configuration
        • Network access and security configuration
        • Basic storage account setup
      • Configure CycleCloud, schedulers, identity, and monitoring
        • Resource orchestration with Azure CycleCloud
        • Job scheduler deployment and migration
        • Identity management and user authentication
        • Monitoring with Azure Monitor and Grafana
        • Cost accounting and budget tracking
      • Configure HPC storage and migrate data
        • Choose the right storage for your HPC workload
        • HPC storage options Lustre, ANF, Blob, Azure Files
        • Migrate data to Azure Data Box, AzCopy, rsync
      • Provision HPC and GPU compute nodes
        • Compute node selection and quota planning
        • VM image configuration Azure HPC images
      • Enable user access and job submission
        • End user entry access
        • User entry access configuration and job submission
• HPC and AI performance and benchmarking
  • HPC and AI benchmarking overview
  • Best practices for HPC workloads and storage performance
  • HPC and AI benchmarking applications reference
  • Optimize performance for HPC and AI virtual machines
  • Run your first benchmark using STREAM
• AI training
  • Choose your platform for AI training and fine-tuning