SCR caches checkpoint data in storage on the compute nodes of a Linux cluster to provide a fast, scalable checkpoint / restart capability for MPI codes.
Many apps don't bother to track whether they have read or written files correctly. Instead they detect an error and immediately abort at the I/O call. Such apps often then just pass True in their calls to scr.complete_restart(valid) and scr.complete_output(valid).
To simplify, this makes the valid parameter optional with a default value of True.
Many apps don't bother to track whether they have read or written files correctly. Instead they detect an error and immediately abort at the I/O call. Such apps often then just pass
Truein their calls toscr.complete_restart(valid)andscr.complete_output(valid).To simplify, this makes the
validparameter optional with a default value ofTrue.