Data Day 2022
Slides and Recordings
You can find all slides as well as tutorial materials on the Data Day 2022 github page. Slides are also linked below in the Downloads section. Video recordings coming soon.
Abstract
NERSC is rebooting its data-centric training, Data Day, as a hybrid two-day event on October 26-27, 2022. Join us for talks, tutorials, and hands on hacking designed to get you up and running with the latest and greatest data-focused tools for scientific computing on Perlmutter.
This event is geared towards HPC users of all experience levels, but will focus on intermediate and advanced topics which will help data workloads run performantly at scale on Perlmutter. First time HPC users may want to attend or review our recent New User Training before attending Data Day. We strongly encourage attendance from Graduate Students, Postdocs and others who have an active role in development. With the announcement of Cori's approaching retirement in January 2023, users migrating their data applications from Cori to Perlmutter are also highly encouraged to join.
Morning sessions will include advanced tutorials on topics such as python, containers, workflow tools, data transfer and more, as well as lightning talks from users on the latest data applications at NERSC. All participants are encouraged to submit a brief lightning talk! Participants will have access to NERSC experts and the opportunity to implement tools in their workloads in afternoon hacking sessions.
Agenda
The event agenda is below:
Date | Start Time (PDT) | Topic | Speaker |
Oct 26 | 10:30 AM | Introduction to Data Day (slides) | |
10:40 AM | Evolution of Data Services for Science | Wahid Bhimji | |
11:00 AM | I/O Profiling on Perlmutter with Darshan | Alberto Chiusole & Jean Luca Bez | |
12:00 PM | Lunch Break | ||
1:00 PM | Containers-as-a-Service: Spin | Cory Snavely | |
1:30 PM | Workflows: Pegasus Workflow Manager | Nicholas Tyler | |
2:00 PM | Superfacility API | Bjoern Enders | |
2:30 PM | Hacking session | ||
4:30 PM | End | ||
Oct 27 | 10:30 AM | Containers for HPC: Shifter and Podman | Daniel Fulton |
11:00 AM | Scaling Python Applications | Daniel Margala | |
11:30 AM | Julia | Johannes Blaschke | |
12:00 PM | Lunch Break | ||
1:00 PM | Data Visualization: Altair Demo | Annette Greiner | |
1:30 PM | Deep Learning at Scale on Perlmutter | Steven Farrell | |
2:00 PM | Python on GPUs: JAX | Nestor Demeure | |
2:30 PM | Hacking session | ||
4:30 PM | End |
Registration
This event has already taken place and registration is closed.
Presentation Materials
- Slides attached below
- Video Recordings
Downloads
-
01-Data-Services-for-Science-Oct-2022.pdf | Adobe Acrobat PDF fileEvolution of Data Services for Science - Wahid Bhimji
-
02a-I-O-profiling-on-PM.pdf | Adobe Acrobat PDF fileI/O Profiling on Perlmutter with Darshan - Part 1 - Alberto Chiusole
-
02b-DXT-Explorer-+-Drishti.pdf | Adobe Acrobat PDF fileI/O Profiling on Perlmutter with Darshan - Part 2 - Jean Luca Bez
-
04-Data-Day-Pegasus.pdf | Adobe Acrobat PDF fileWorkflows: Pegasus Workflow Manager - Nicholas Tyler
-
06-Containers-for-HPC-Shifter-and-Podman.pdf | Adobe Acrobat PDF fileContainers for HPC: Shifter and Podman - Daniel Fulton
-
07-Scaling-Python-Applications.pdf | Adobe Acrobat PDF fileScaling Python Applications - Daniel Margala
-
09-Python-Vis-with-Altair.pdf | Adobe Acrobat PDF fileDeclarative Python Visualization With Altair - Annette Greiner
-
10-Farrell-DataDay22-DeepLearning.pdf | Adobe Acrobat PDF fileDeep Learning at Scale on Perlmutter - Steven Farrell