An Application-Level Incremental Checkpointing Mechanism with Automatic Parameter Tuning

Takizawa, Hiroyuki; Amrizal, Muhammad Alfian; Komatsu, Kazuhiko; Egawa, Ryusuke

doi:10.1109/candar.2017.96

Cited by 3 publications

(1 citation statement)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The alternative to system-level checkpointing is application-level checkpointing, where it is the application whom writes its own state into a checkpoint file, and reads it on the restart stage. While this requires explicit code inside the application and hence is no longer transparent, it gives the application the ability to decide when checkpoints should be taken and what should be contained in the checkpoint (see for examples works on incremental checkpointing applied to application-level [20,21]).…”

Section: Status Of Checkpoint/restart Librariesmentioning

confidence: 99%

Job migration in HPC clusters by means of checkpoint/restart

et al. 2019

View full text Add to dashboard Cite

Until now, jobs running on HPC clusters were tied to the node where their execution started. We have removed that limitation by integrating a user level Checkpoint/Restart library into a Resource Manager, fully transparent to both the user and running application. This opens the door to a whole new set of tools and scheduling possibilities based on the fact that jobs can be migrated, checkpointed and restarted on a different place or in a different moment, while providing fault-tolerance for every job running on the cluster. This is of utmost importance in the future generation of exascale HPC clusters, where an increasing degree and complexities of efficient scheduling make it challenging to obtain the required degree of parallelism demanded by the applications.

show abstract

Section: Status Of Checkpoint/restart Librariesmentioning

confidence: 99%