Quantum #70

Issue #70 of the weekly HPC newsletter by HMx Labs. We finally update our website and figure out what we can do for you, cloud HPC release notes and some talk of benchmarking.

Quantum #70

It’s taken us a lot longer than I thought it would, but we have finally removed that little beta sticker from the top of the HMx Labs website! 🎉 What this means is that we’ve finally figured out how we’re trying to position ourselves in the market and what it is that we can do for you. While this has been relatively clear to me in the HPC space for a while (in case you hadn’t noticed) it was far from clear to us in the non HPC side of the business what its core focus should be.

We think we’ve nailed it now though. Data.

Not only do have a long history of working on open source data management (Waltz) but we were also responsible for building out a data mesh before the term was even coined. With the recent explosion in interest around AI it felt natural we focused on the two things that enable it. Compute and Data. At scale.

We’re still missing some content, notably case studies but we hope to have them out there soon.

And of course what happened in the rest of the world of HPC is below.


In The News

Updates from the big three clouds on all things HPC.

HPC Cloud Updates WE 27 Jul 2025
Updates to AWS, Azure & GCP in the last week relevant for HPC practitioners. A few new AWS instance types, some improvements to Cluster Director in Google Cloud and fun with AI and Slurm on Azure

I’m not sure how willing people will be to share large scale HPC/A clusters, if they do the network will certainly be one problem to solve … We still have solved shared memory bandwidth and CPU caches on cloud though so I won’t hold my breath

Multi-Tenant HPC and AI: How the Network Can Make or Break the System
As artificial intelligence (AI) and HPC applications rapidly evolve and diversify, organizations are rethinking how to design and scale the infrastructure that powers these systems. Central to this transformation is […]

Some insight into what’s going on over at Intel.

Intel Puts The Process Horse Back In Front Of The Foundry Cart
It is beginning to look like Intel plans to milk the impending 18A manufacturing process for a long time. The 18A process is akin to a refined 2 nanometer

From HMx Labs

My thoughts this week were all around benchmarks, first up, how do you measure performance of your HPC cluster as a whole? I don’t have an answer, but I was looking for input on what parts matter the most to everyone else

Measuring Cluster Performance
Attempting to come up with a realistic distributed benchmark that includes not only the hardware but the also the software such as the scheduler, the data plane and internode communication

And of course if you’re measuring performance… you should know not only what you’re measuring but why you’re measuring it

The Bi-Modal Nature of Benchmarks
Benchmarks can serve two subtly different purposes, are you sure you’re getting what you need from yours?

Know someone else who might like to read this newsletter? Forward this on to them or even better, ask them to sign up here: https://cloudhpc.news