HPC Cloud Updates WE 03 Nov 2024

Updates to AWS, Azure & GCP in the last week relevant for HPC practitioners

HPC Cloud Updates WE 03 Nov 2024

AWS

Need to validate your EC2 scale out was successful before before using it? New auto scaling gives you final validation

Amazon EC2 Auto Scaling supports final validation time for instance refresh - AWS
Discover more about what’s new at AWS with Amazon EC2 Auto Scaling supports final validation time for instance refresh

EC2 m8g instances now available in Dallas

Introducing Amazon EC2 M8g instances in Dallas Local Zone - AWS
Discover more about what’s new at AWS with Introducing Amazon EC2 M8g instances in Dallas Local Zone

Having trouble nailing your IO latencies? CloudWatch for EBS might help

New Amazon CloudWatch metrics for monitoring I/O latency of Amazon EBS volumes - AWS
Discover more about what’s new at AWS with New Amazon CloudWatch metrics for monitoring I/O latency of Amazon EBS volumes

r7gd now in available in Jakarta, UAE and GovCloud

Amazon EC2 R7gd instances now available in additional regions - AWS
Discover more about what’s new at AWS with Amazon EC2 R7gd instances now available in additional regions

Azure

Resize your SSDs without stopping the instance

Generally Available: Live Resize for Azure Premium SSD v2 and Ultra Disks
Generally Available: Live Resize for Azure Premium SSD v2 and Ultra Disks

and those SSDs in v2 form are now available in more places too

Generally Available: Azure Premium SSD v2 Disk Storage is now available in more regions
Generally Available: Azure Premium SSD v2 Disk Storage is now available in more regions

Google Cloud

Google copies the automotive world and coins the term hypercomputer. And throws in AI for good measure. Yes, they invented the AI Hypercomputer. 

Except its not a computer. It’s a collection of Google Cloud tech to let you build your own large scale compute. Would the marketers at Google please step away from the AI Hype Machine (patent pending).

I wonder if there’s a (negative) correlation between the amount of hype foisted upon us and the actual utility of the product.

Trillium sixth-generation TPU is in preview | Google Cloud Blog
AI Hypercomputer can now include Trillium TPUs in preview, Axion-powered C4A VMs in GA, and coming soon, A3 Ultra VMs powered by NVIDIA H200 GPUs.

Axion (ARM) powered c4a compute instances now generally available)

Try C4A, the first Google Axion Processor | Google Cloud Blog
The custom Arm-based processor is designed for general-purpose workloads like web and app servers, databases, analytics, CPU-based AI, and more.