The goal of this talk is to provide insight on how Boston University (BU) supports GIS workflows that use GIS open source software on our High Performance Computing Cluster and provide examples of what kinds of research is being done using these workflows.

I am a scientific programmer/analyst at BU and I am part of Research Computing Services. My group maintains and supports our high performance cluster (HPC). I am part of the Applications Team and my group provides onboarding services and training for researchers, students, and staff at BU, and my specific area of expertise is in GIS computing and workflow development.

BU's HPC resources is available to use by any community member of our 17 colleges and schools. Our cluster runs an Alma8 Linux operating system and uses a customized Sun Grid Engine as the scheduler. We have approximately 7,000 processes, 236 GPUS, two petabytes of research data, and over 2,500 users. The cluster has over 1,300 software packages installed to support research and approximately 50 software packages are associated with GIS workflows, with majority of them being open source software, libraries, or standards. This includes GDAL, GrassGIS, QGIS, PostGIS, NetCDF, Zarr, Orfeo Toolbox, Climate Data Operators (CDO), GEOS, PROJ, PDAL, and R and Python GIS libraries. Our HPC system supports running jobs as batch jobs, but also interactive jobs, so researchers can map their GIS data using programs like QGIS, or plotting libraries from R and Python.

As part of my role, I provide training and consulting on how to use these open source software on our cluster. I also provide support on how to best optimize their GIS workflow to take advantage of parallelization and also how to process very large datasets. With supporting 17 colleges and schools, I see the different ways opens source geospatial software is used at the university. Some example research projects include:
- Center for Remote Sensing running phenology change analysis of the entire continental US for span of a decade.
- Public health projects that model which communities are impacted by PM2.5 air pollutants.
- Biology department tracking monarch butterfly migrations.
- Dental School project identifying service area gaps coverage of dental offices in the state of Massachusetts.

For my talk I will cover the following three items:

1.) Provide a basic understanding of what is a high performance cluster and how do our researchers interact with it.
2.) An overview of what types of open source software for geospatial is used on our cluster.
3.) Provide examples of research that is done using open source software for geospatial on our cluster.