HPC Cloud Updates WE 27 Apr 2025

Updates to AWS, Azure & GCP in the last week relevant for HPC practitioners. New VM types from AWS and Azure and a pretty exciting enhancement to VM Scale Sets.

HPC Cloud Updates WE 27 Apr 2025

AWS

Optimise Ansys running on AWS with Ansys Gateway (AWS Marketplace offering)

Optimizing HPC workflows with automatically scaling clusters in Ansys Gateway powered by AWS | Amazon Web Services
Ansys Gateway powered by AWS now has an integration with AWS ParallelCluster to enable users deploy on-demand HPC clusters for running Ansys simulations on AWS. This allows engineers to run large-scale simulations efficiently while optimizing cloud costs by dynamically adjusting resources based on simulation workload requirements. In this blog post, we describe the architecture, workflow, and Amazon EC2 recommendations for running Ansys applications in Ansys Gateway.

New AZ planned for the N. Virginia region

In the works – New Availability Zone in Maryland for US East (Northern Virginia) Region | Amazon Web Services
AWS is adding a new Availability Zone in Maryland to the US East (Northern Virginia) Region by 2026, enhancing redundancy and supporting growing AI workloads while maintaining high-bandwidth, low-latency network connections.

Changes to the Carbon Footprint Tool should help see what how many polar bears your HPC workloads are making homeless

Customer Carbon Footprint Tool has new features and an updated methodology - AWS
Discover more about what’s new at AWS with Customer Carbon Footprint Tool has new features and an updated methodology

New Graviton4 instances incoming! c8g, m8g and r8g instance types now GA in Ohio, N. Virginia and Oregon.

Introducing Amazon EC2 C8gd, M8gd, and R8gd instances - AWS
Discover more about what’s new at AWS with Introducing Amazon EC2 C8gd, M8gd, and R8gd instances

New Instances: i4g in Sydney, m8g in N. California, r7gd in London, Hyderabad and Osaka, c7gd in Canada (Central), and London, c6id in Paris

Amazon EC2 I4g instances are now available in AWS Asia Pacific (Sydney) Region - AWS
Discover more about what’s new at AWS with Amazon EC2 I4g instances are now available in AWS Asia Pacific (Sydney) Region
Amazon EC2 M8g instances now available in AWS US West (N. California) Region - AWS
Discover more about what’s new at AWS with Amazon EC2 M8g instances now available in AWS US West (N. California) Region
Amazon EC2 R7gd instances are now available in additional AWS Regions - AWS
Discover more about what’s new at AWS with Amazon EC2 R7gd instances are now available in additional AWS Regions
Amazon EC2 C7gd instances are now available in additional AWS Regions - AWS
Discover more about what’s new at AWS with Amazon EC2 C7gd instances are now available in additional AWS Regions
Amazon EC2 C6id instances are now available in AWS Europe (Paris) region - AWS
Discover more about what’s new at AWS with Amazon EC2 C6id instances are now available in AWS Europe (Paris) region

Azure

New NVads_v5 VM family powered by AMD 4th Gen EPYC CPUs and AMD Radeon Pro v710 GPUs

Introducing NVads V710 v5 series VMs | Microsoft Community Hub
Cost-optimized AI inference, virtual workstations, and cloud gaming. AI inferencing and graphics-intensive applications continue to demand cost-effective,…

I’ve often complained we don’t pay enough attention to monitoring and visualisation of HPC workloads, here’s an counterpoint. Somehow though I spent more time wondering if that title picture is AI generated or not (I think it must be) or if that’s some fancy new kind of office/data centre that Microsoft has. Must be noisy in there!

Monitoring HPC & AI Workloads on Azure H/N VMs Using Telegraf and Azure Monitor (GPU & InfiniBand) | Microsoft Community Hub
As HPC & AI workloads continue to scale in complexity and performance demands, ensuring visibility into the underlying infrastructure becomes critical.…

Guess that ties in nicely with this new feature in preview too, Metrics usage Insights

Azure updates | Microsoft Azure
Subscribe to Microsoft Azure today for service updates, all in one place. Check out the new Cloud Platform roadmap to see our latest product plans.

This looks very interesting for scaling your compute: Instance Mix for VM Scale Sets. Let’s you mix different VM types in the same Scale Set.

Azure updates | Microsoft Azure
Subscribe to Microsoft Azure today for service updates, all in one place. Check out the new Cloud Platform roadmap to see our latest product plans.

The documentation give more details and if you were thinking of using this to upgrade to the latest generation of VM type but topping up capacity with an older generation then that’s even a suggested use case 😄

Use multiple Virtual Machine sizes with instance mix - Azure Virtual Machine Scale Sets
Use multiple Virtual Machine sizes in a scale set using instance mix. Optimize deployments using allocation strategies.

There even appears to be a lowest price allocation strategy to automatically deploy the cheapest compute… I wonder if that accounts for performance adjusted pricing. I’m definitely going to have to play with this a bit.

Microsoft’s own blog about it here:

General Availability: Instance Mix for Virtual Machine Scale Sets with Flexible Orchestration Mode
Today we’re announcing the General Availability of Instance Mix on Virtual Machine Scale Sets with Flexible Orchestration Mode, enabling customers to mix &…

My biggest question though, and whilst I’d expect this from AWS, its unusual for Azure: What’s the direction here? Should I be using VM Scale Sets or Azure Compute Fleet? There seems to be a lot of overlap between the two.


Google Cloud

Nothing to see here. Move along folks.