Here are new options!
***************************************************************************
Actually, there is a company, Platform Computing, which provides a
checkpoint-restart option in its Load Sharing Facility (LSF) software.
It is necessary to compile and link the program (Fortran or C) with a
special library and to submit it with the appropriate options to the
LSF batch system. It works very well and is especially useful in
migrating running jobs from one (loaded) machine to another, empty
machine. On the other hand, it is not free and it is not inexpensive.
Peter Stern
****************************************************************************
*
If you want a piece of software that does check-pointing automatically
(assuming you are running code written in C), you can try "condor"
(it is a distributed/migration system that aloows a job to be migrated
from machine to machine during execution....it does so by check-pointing)...
juts search for it on the web.
Charles C. H. Jui
Received on Wed Sep 02 1998 - 20:43:07 NZST