Introduce yourself for the HPC workshop!

Many of us are traveling far and wide to attend this event. Take a moment to say hello, where you’re coming from, and talk a bit about what the world looks like from your perspective! Bonus points it you can work in how your institution uses (or might use) jupyter in its work!

I can start -

My name is Chris Holdgraf, I work with the Jupyter team here at Berkeley, as well as with the Berkeley Institute for Data Science (BIDS) and the Division of Data Science. We help researchers and educators do their jobs more effectively by building tools + infrastructure for interactive computing. We help the university run a JupyterHub for several classes, including a large 1,400 student introduction to data science course. I’m excited to see what everyone has been up to at their institutions, and ways that the Jupyter Project can better-support people at shared facilities!

My name is Dan Allan. I work at Brookhaven National Lab in New York. The lab’s new synchrotron, NSLS-II, is a user facility that generates data across many science domains and techniques with widely varying scales, resource requirements, and software sophistication. NSLS-II relies on Jupyter as important way for users to access, explore, and process their data. We are focused on notebook sharing models—fast on-boarding for new users is an important part of our mission—and on making HPC resources more accessible to more users.

1 Like

Hello All, my name is Maksim Rakitin, I work at NSLS-II, Brookhaven National Lab, Upton, NY with Dan Allan and other bright scientists and engineers in a group called “Data Acquisition, Management and Analysis” (DAMA). As Dan mentioned, we use Jupyter Notebook/Lab for data analysis (and sometimes for data acquisition) at scale. We also heavily use mybinder/Jupyter for our tutorials and various examples of how our libraries can be used. I hope the workshop will be a great opportunity to exchange the experience between the DOE labs, academia and industry. Hope to meet you all next week!

1 Like


My name is Teddy Rendahl. I work at SLAC National Accelerator Laboratory in the “Photon Controls and Data Systems” group. We currently use Jupyter as a way for users to analyze the data they collect after using our facility. I’m specifically interested in expanding the use of notebooks to be a main tool during data collection, allowing users to capture and share experimental techniques. This will both simplify downstream analysis while ensuring we perform more repeatable scientific experiments.

I’m Rob Nagler from RadiaSoft. We are an accelerator and beam physics consultancy. Many of our staff rely on Jupyter for their daily work, which consists of developing models and validating them. We also provide support to workshops and classes in the accelerator physics community by letting them use our JupyterHub cluster. is a open and free for use by the public, and contains many popular accelerator and beam physics codes as well as other popular tools such as keras, tensorflow, etc.

Our use of Jupyter has evolved over the last few years to have a common shared cluster with public and private server pools. All of our codes use MPI so we also allocate clusters to staff, who tend to run jobs over days and weeks. We spend a lot of time optimizing our Docker builds, because the codes and support libraries are rather large, and contain much legacy.

I’m very interested to learn solutions to resource management, legacy builds, sharing notebooks, user customizations, running MPI codes, and security issues.

I’m Dan Katz. I work at the University of Illinois at Urbana-Champaign. My focus is on sustainable software, including good/best practices for software citation and software-oriented career paths, particularly in academia. I also am part of the Parsl ( team, and work on a number of projects that try to develop communities in software-focused areas. See for more info. I’m here because I want to talk about Parsl (which allows parallelizing Python programs, including building workflows) and how it works in Jupyter, and to find out if there are things that others are doing we can use, or if there are things we can work on together, such as how we describe HPC systems to let tools submit jobs to them, and how we describe applications that we might want to program on top of.

Shreyas Cholia - Usable Software Systems Group (Computational Research Division, LBL) and NERSC. We have a number of engagements with science groups to enable Jupyter based tools for scientific workflows. A big focus of our work involves enabling interactive human-in-the-loop computing at scale. We are interested in using Jupyter to orchestrate distributed tasks and workflows. I also help with Jupyterhub deployments at NERSC and for a couple of individual science projects.

I’m Bill Riehl. I work at LBNL on the DoE Systems Biology Knowledgebase (KBase), which enables biologists to analyze, share, and build narratives from their data. We use the Jupyter Notebook as the main interface to our job running and data management systems. I’m interested in updating to Jupyterlab as our next generation interface, and making use of other Jupyter-based technology to link to HPC systems, including NERSC.

I’m Russell Neches. I’m a postdoc in the Prokaryotic Super Program at the Joint Genome Institute (we’ll be here at the main LBNL campus once they finish building our building). I mostly work on microbial evolution and ecology with graph theory and machine learning.