eehindero announces the public preview of Azure VM Customization, highlighting new support for disabling SMT/hyperthreading and setting constrained core counts, which empowers users to optimize performance and licensing costs.

Public Preview: VM Customization in Azure Enables Disabling Multithreading and Constrained Cores

Author: eehindero
Published: Oct 17, 2025

Azure has introduced a new public preview feature called VM Customization, which adds two powerful options:

  • Disable Simultaneous Multi-Threading (SMT/HT): Allows supported VMs to run with one thread per core, granting workloads exclusive access to physical cores. This is ideal for latency-sensitive or performance-critical workloads.
  • Constrained Cores: Lets users select custom vCPU counts for each VM size, tailored to licensing needs or workload demands, without altering memory, storage, or I/O resources.

Key Benefits

  • Performance Optimization: Disable hyperthreading for full-core isolation, which can improve performance consistency for workloads like high-performance computing (HPC), database servers, analytics, and more.
  • Software Licensing Cost Control: Reduce licensing expenses by deploying high-memory or high-bandwidth VMs with fewer active cores. Especially beneficial for SQL Server, Oracle, SAP, and other BYOL/per-core licensing models.
  • Flexibility: Match VM resources to specific workload profiles or licensing requirements while maintaining all other VM capabilities.
  • Compliance: Simplifies adherence to software compliance models that require precise core counts.

Availability

  • Currently in public preview in regions: West Central US, North Europe, East Asia, and UK South.
  • Accessible via:
    • Azure Portal
    • ARM Templates
    • Azure CLI
    • PowerShell
  • Only first-party OS images are supported; marketplace images with third-party licensing are not supported.

How to Join the Preview

To participate, fill out the survey form here.

Example Scenarios

  • Database Optimization: Deploy a SQL Server VM with high memory and I/O bandwidth but only the exact number of vCPUs matching license entitlements.
  • High-Performance Computing: Assign physical cores (with SMT/HT off) to improve consistency and reduce latency.
  • Cost Optimization for Enterprises: Utilize constrained core counts to avoid over-provisioning and save on licensing costs.

Additional Notes

  • These features deliver deeper customization for various enterprise workloads in Azure, empowering technical teams to optimize infrastructure and budgets without compromising on capability.
  • More information can be found in the Azure Compute Blog.

For detailed setup instructions and supported VM series, visit the official Azure documentation or follow the Azure Compute Blog for updates.

This post appeared first on “Microsoft Tech Community”. Read the entire article here