[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Condor-users] lightweight cluster/grid engine



I wrote a very lightweight cluster/grid engine for dedicated networks,
mostly because condor didn't allow me to "subqueue" other jobs, but
also because I thought it was condor doing far more than I needed for
my dedicated compute cores.

It works OK.  I still use condor for the main cluster and for shared
machines and other stuff like that, but I like my perl-engine (faster
execution time, less overhead, fewer features, slightly better average
throughput).

I then "if_full_exec" to the main cluster via "condor_run" when the
mini cluster is full.

Is anyone interested in looking at it? Or in telling me that there
already is one and I should stop.

- Erik