High Performance Computing
Exploit the power and elasticity of the cloud combined with the user friendliness of HPCMinds to run your high demanding problems as fast as possible or at a minimum cost.
WHAT WE DO?
HPC minds is an easy to use platform which enables you to efficiently run your models with high computational requirements on cloud infrastructure. Our platform will allow you to focus completely on the creation and fine-tuning of your model while we take over its execution and ensure that it is run on the best resource combination available.
HOW DOES IT WORK?
When you upload your model to our platform, our optimization engine scans through thousands of possible resource allocation options and selects the best one.
This includes many different types of cloud instances as well as different numbers of them. You can think of this as a “Skyscanner” but for cloud computing.
The engine estimates the costs of different resource allocation strategies thanks to a data driven approach which learns from past experience. This means that as you use our platform, our engine also learns to schedule your individual models better and adapts to your needs.
After your model has been uploaded, you are given a choice between different resource configurations that were identified by our engine. These include a very high-speed execution, a low-price option as well as a tradeoff of the two options.
Once you select your desired configuration, we will automatically provision the necessary resources and execute your model on them. You will receive a message notifying you when your results are ready for you to download!
Parallelizing CFD Simulations
NewsSelecting the correct level of parallelization for CFD simulations
In this blog, we will look at how to select the level of parallelization for a CFD simulation on the cloud. To do so we will use the example of the Weather Research and Forecast (WRF) model, which is a type of Computational Fluid Dynamics (CFD) simulation. Our experiments will be run using c6i and c5d instances from Amazon Web Services (AWS). But first let’s start with some background on how a CFD simulation is parallelized.