GeoDict User Meeting 2023
Cloud capabilities in GeoDict 2023
Our future is digital and working remotely from home has become a new normal for us. And even from home we want to access and orchestrate computational resources that fit our current needs and are expandable limitlessly to match peak usage in demanding projects. Therefore, companies are increasingly outsourcing their IT infrastructure by moving from on-premise hardware towards cloud-based solutions. In addition, these have the advantages of reduced maintenance cost and lower carbon footprint. This is particularly obvious when considering computation-intensive simulation applications such as GeoDict.
In this presentation, I will outline our solutions for remote and/or cloud usage of GeoDict 2023. In particular, I will demonstrate the two different solutions offered by Math2Market for executing GeoDict simulations in the cloud:
The Math2Market Cloud provides a web-interface to launch virtual workstations which can be accessed directly within the web-browser, providing a seamless experience of using GeoDict interactively on a powerful machine. Input data such as CT scans can be uploaded to the machine, simulations can be run unattended, and the results can be downloaded to your computer. Your local machine does not need to be a high-performance computer and does not require an installation of GeoDict.
The GeoDict Cloud follows a different philosophy in that it is integrated into your local GeoDict GUI. Almost every operation that you can do in GeoDict, such as running a flow simulation using FlowDict or generating a grain structure using GrainGeo can be delegated to the cloud with the simple press of a button. It’s as easy as clicking “Run in Cloud” instead of “Run”. Using this feature, any number of simulations can be started in quick succession and will execute simultaneously on individual virtual machines. The new Cloud Project Manager allows you to keep track of your running simulations and download the results of any finished cases.
Via GeoPy, whole parameter studies can be started with a few clicks, running all individual cases in parallel on individual machines. The results of all cases in your study will be available in the time that a single case would need to execute on your local machine.
Both cloud solutions provide machines from a single core up to hundreds of cores with terabytes of RAM to fulfill even the most demanding simulation and image processing requirements.