Checkpoint/Restart Mechanism

• Blocking Co-ordinated Checkpoint

  • State of chares are checkpointed to disk
  • Collective call MPI_Checkpoint(DIRNAME)

• The entire job is restarted

  • Virtualization allows restarting on different # of Pes
  • Runtime option
    • > ./charmrun pgm +p4 +vp16 +restart DIRNAME

• Simple but effective for common cases