| Note: | Before you consider using the Checkpoint/Restart function refer to the LoadL.README file in /usr/lpp/LoadL/READMES for information on availability and support of this function. |
Checkpointing is a method of saving the state of a job so that if the job does not complete it can be restarted from the saved state rather than starting the job from the beginning. Both serial and parallel jobs can be checkpointed. LoadLeveler provides mechanisms for a program to checkpoint itself as well as providing means for checkpoints to take place outside of the programs control.
For more information see Step 14: Enable checkpointing.