of the daemon rather than restoring the bitmap since the nodes in a system
    can change (be added or removed).

    request and job_details structure in slurmctld.

    within the job step credential.
svn merge -r15520:15530 https://eris.llnl.gov/svn/slurm/branches/multi-core

    killed even if the node where srun executes goes down (unless the
    --no-kill option is used, previous termination logic would fail if 
    srun was not responding).

    specific subsystems. This is just the infrastructure, flag is unused
    right now (work on tomorrow).

    (cons_res_preempt.patch from Chris Holmes, HP).

    the TaskPluginParam configuration parameter rather than slurmd's
    SLURM_ENFORCED_CPU_BIND environment variable.

  (also added output of SacctDefaultCommand from "scontrol show config")

  and Moe Jette. There are still a couple of end-cases that need to be fixed. We also
  need to push the core identification out to the slurmd for task binding. sched/gang
  needs revision to support this too.

    a job's completion before allocating already released resources to pending
    jobs. This can be used to reduce fragmentation of resources.