Cloud-powered PDC Computations for a Runestone Interactive Textbook
Collaborative Undergraduate Research and Inquiry 2021
Abstract
Runestone textbook is written for beginning programmers who want to explore the concept of PDC. The ActiveCode, a feature of Runestone, provides learners a chance to interact with their own PDC code by compiling and analyzing the result. Kubernetes and Docker are the key technologies to provide the portability, extensibility, sustainability, and management of Runestone’s backend. By receiving the request from the user through ActiveCode, we then send these requests to the backend where pods are ready to process the compilation. We then send back the result to the Runestone website interface for the user to analyze the result of their own code. Our service includes MPI and OpenMP requests. By containerizing the whole process using Kubernetes and Docker, we can modify the Docker images, the template file, etc. to different needs. Our future plan is to provide more containers for mpi4py, CUDA and other PDC computations.
Authors
Dr. Richard A. Brown, Khang V. Huynh, Tanaka Khondowe, Geroge Kokalas
Grants
NSF grant: DUE-1822480 (CSinParallel)
Poster Presentation
Collaborative Undergraduate Research and Inquiry 2021 Closing Symposium, St. Olaf College, Minnesota, United States