On Mar 26, 2009, at 12:17 AM, Christoph Spiel wrote: > Benjamin - > > On Wed, Mar 25, 2009 at 01:50:58PM -0700, Benjamin Ylvisaker wrote: >> 2) In C or bash or whatever, implement basically the same thing, but >> externally at the process level. >> >> Has anyone else dealt with this issue? Any ideas about what would be >> best? > > There are several load-levelling systems > out there, which distribute processes among a > cluster of computers. > > I'm aware e.g. of Condor > http://www.cs.wisc.edu/condor/ > which features "periodic_remove". This control > allows to kick off a job from the queue if > certain (user-definable) critera are met, > e.g. cpu time or wall-clock time exceeded. > > HTH! > > /Chris >
My group is in fact using Condor, but I thought it might be nice to implement a limit that didn't depend on running in the Condor context, since we sometimes run jobs directly on machines, instead of through Condor. I ended up putting a simple periodic polling mechanism into my optimization, which seems to do the job well enough. No need for extra threads or processes. Thanks, Ben ------------------------------------------------------------------------------ _______________________________________________ CIL-users mailing list CIL-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/cil-users
