07-11, 15:45–15:55 (Europe/Amsterdam), While Loop (4.2)
Fugro is a global leader in geospatial data analytics, and as a part of our product workflows we need to deal with large amounts of data as efficiently as possible. In this presentation, we will describe how we use Julia to accelerate our HPC applications and how we integrate this with Amazon Web Services to distribute computing on the cloud. We will also cover some of the challenges and the insights we’ve gained on working with Julia in industry.
Here at Fugro, we process large amounts of geospatial data in numerous HPC workflows to provide our clients with insights about real-world assets. Efficient processing enables faster delivery times for clients and reduced costs. We write our applications in Julia and have built systems using Github Actions that can seamlessly integrate these applications with cloud infrastructure hosted by Amazon Web Services (AWS).
In this talk, we will outline our process at Fugro to develop a production-ready HPC application with Julia. This will start with how we build these packages into binary system images and integrate them into docker images. We will then show how to use AWS and Github to deploy these docker images to a cloud environment, as well as some design considerations along the way to ensure you get the most out of your application. Throughout we will share our experiences and lessons learnt for anyone starting or considering using Julia in an industrial application.
Mathematician and data scientist, interested in using Julia for high performance computing and cloud workflows
Dr. Megan Dawson is the Modelling Automation Team lead at Fugro, responsible for numerous cloud based HPC workflows efficiently processing large amounts of geo-data. With 15 years experience in industry her passion is problem solving to developing practical (and production-able) solutions.