BEGIN:VCALENDAR
VERSION:2.0
METHOD:PUBLISH
CALSCALE:GREGORIAN
PRODID:-//WordPress - MECv6.2.5//EN
X-ORIGINAL-URL:https://www.rc.fas.harvard.edu/
X-WR-CALNAME:FAS Research Computing
X-WR-CALDESC:.HARVARD UNIVERSITY | FACULTY OF ARTS & SCIENCES
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-PUBLISHED-TTL:PT1H
X-MS-OLK-FORCEINSPECTOROPEN:TRUE
BEGIN:VEVENT
CLASS:PUBLIC
UID:MEC-74a9013d7b74ca4a87c9f6ae51956d66@rc.fas.harvard.edu
DTSTART:20261105T180000Z
DTEND:20261105T193000Z
DTSTAMP:20261026T214900Z
CREATED:20261026
LAST-MODIFIED:20261027
PRIORITY:5
TRANSP:OPAQUE
SUMMARY:Save Your Progress: Checkpointing On FASRC Clusters (virtual)
DESCRIPTION:Save Your Progress: Checkpointing On FASRC Clusters (virtual)\n \nDescription: When a cluster job runs out of time or stops unexpectedly, you can lose hours or days of computation. Learn to save your calculation’s progress and resume from the last saved point instead of starting over. Checkpointing allows a running program to periodically save its computational state so that, if interrupted, it can resume from the last saved point rather than starting over. Similar to saving a document while working, checkpointing protects your computational progress and helps avoid wasting valuable computing time and resources.\nIn this virtual training session, participants will learn practical techniques for implementing checkpointing and restarting computational jobs on FASRC clusters. We will explore different checkpointing approaches, including application-level checkpointing, transparent checkpointing without modifying source code, and checkpointing capabilities built into scientific applications and machine-learning frameworks.\nHands-on Examples\n\nResume a Python or C calculation after an interruption.\nContinue a GROMACS molecular simulation across multiple jobs.\nResume PyTorch model training from a saved checkpoint.\nSave and resume some programs you didn’t write, including parallel (MPI) applications, without adding checkpoint code.\n\n \nYou’ll also learn to save progress before a job’s time limit, automate restarts, and use FASRC’s requeue partitions.\nNo prior experience with checkpointing is required.\n \nWho should attend?\nResearchers running long calculations, simulations, or model training on FASRC clusters\n \nDate: Thursday, Nov 5, 2026\nTime: 1:00 – 2:30 PM \nLevel: Intermediate \nCost: Free\nPresenter: Plamen Krastev \nFacilitators: Nathan Weeks, Brad Spitzbart, & Manasvita Joshi\nPrerequisites: \n\nFASRC cluster account. If you do not have an account, see Request a FAS Research Computing Account well in advance ( https://docs.rc.fas.harvard.edu/kb/how-do-i-get-a-research-computing-account/ ) \nPrevious experience submitting batch jobs on a FASRC cluster\nBasic familiarity with Slurm job scripts\nBasic programming experience in Python, C/C++, or Fortran. Participants interested in the AI/ML examples should have basic familiarity with at least one supported machine-learning framework\n\nRegistration: Register on Zoom for this session: REGISTER ( https://harvard.zoom.us/meeting/register/REtJQ_ghTB-rcV13E9mE8Q )\nAfter registering you will receive a link to the Zoom meeting for that date.\n
URL:https://www.rc.fas.harvard.edu/events/checkpointing-and-restarting-computational-workflows-workshop-virtual/
CATEGORIES:Training
END:VEVENT
END:VCALENDAR