36185ad964
The new component fixes a number of problems with the old component. The core algorithm is the same, but by changing the data strucutres a bit we have improved performance and memory utilization. There are still a couple corner cases that still need some work. However, I did not want to delay bringing this into the trunk (and v1.3 branch) for too much longer. This commit was SVN r19537.
44 строки
866 B
Plaintext
44 строки
866 B
Plaintext
#
|
|
# An Aggregate MCA Parameter Set to enable checkpoint/restart capabilities
|
|
# for a job.
|
|
#
|
|
# Usage:
|
|
# shell$ mpirun -am ft-enable-cr ./app
|
|
#
|
|
|
|
#
|
|
# OPAL Parameters
|
|
# - Turn off OPAL only checkpointing
|
|
# - Select only checkpoint ready components
|
|
# - Enable Additional FT infrastructure
|
|
# - Auto-select OPAL CRS component
|
|
# - If available, use the FT Thread (Default)
|
|
#
|
|
opal_cr_allow_opal_only=0
|
|
mca_base_component_distill_checkpoint_ready=1
|
|
ft_cr_enabled=1
|
|
crs=
|
|
opal_cr_use_thread=1
|
|
|
|
#
|
|
# ORTE Parameters
|
|
# - Wrap the RML
|
|
# - Use the 'full' Snapshot Coordinator
|
|
#
|
|
rml_wrapper=ftrm
|
|
snapc=full
|
|
#filem=rsh
|
|
|
|
#
|
|
# OMPI Parameters
|
|
# - Wrap the PML
|
|
# - Use a Bookmark Exchange Fully Coordinated Checkpoint/Restart Coordination Protocol
|
|
#
|
|
pml_wrapper=crcpw
|
|
crcp=bkmrk
|
|
|
|
#
|
|
# Temporary fix to force the event engine to use poll to behave well with BLCR
|
|
#
|
|
opal_event_include=poll
|