Performance Normalised Cloud VM Pricing for CFD

We’ve previously shown how to optimise cloud HPC costs through the use of automated benchmarking and price capture in financial services. Let’s see if it works for CFD too…

Performance Normalised Cloud VM Pricing for CFD

Sometimes keeping up with, and sharing, everything we’re up to is a full time job in and of itself. I swear I need a bigger hat room!

Around the start of this year we started working with Cygnas to apply some of our work for HPC in the financial services space to various engineering disciplines. If we can substantially reduce the cost of running financial risk models, why not the cost of CFD models too.

As much as I liked this new hat, finding the time to wear it wasn’t easy. This week, HMx Labs took on its first work experience placement. Well, why not throw him in at the deep end I thought!

So poor unsuspecting Yusuf, found himself on a Monday morning having to suddenly understand cloud (AWS, Azure and GCP at a minimum), HPC, aerodynamics and CFD. Good thing he’s interested in going to university to study Aeronautics huh!

He’s making great progress so far and I’m hopeful we will recreate the analysis we performed last year using COREx with OpenFOAM HPC Bike instead. If we get time, we might add another CFD use case into the mix too.

HPC on Cloud for Financial Risk — A Comparative Analysis
This article presents the results of running the COREx benchmark on over 650 different VM types across the three major cloud providers. It attempts to contextualize these results and help HPC practitioners interpret them to optimise their systems, as well as discuss the methodology employed.

Watch this space if you want to know the answer to which cloud VM is most cost effective for funning CFD! Hit the subscribe button!