Add a new GRES function to return the count of elements of a specific
  GRES name
Add a new GRES funtion to reset environment variables based upon the
  GRES which are usable from the CPUs to which a task is bound

Revert "Change cpu_time measurements from uint32_t to double. This gives more accurate cpu utilization when sampling at 1 second rate. Otherwise cpu utilization will be only 100 or 0. With this change it can be any value between 0 and 100."

This reverts commit eaf8e1a2.

Conflicts:
	src/plugins/switch/cray/util.c

switch/cray: Refine logic to set PMI_CRAY_NO_SMP_ENV environment variable.
Rather than testing for the task distribution option, test the actual
task IDs to see fi they are monotonically increasing across all nodes.
Based upon idea from Brian Gilmer (Cray).

Improve cpu profiling values

Change cpu_time measurements from uint32_t to double. This gives more accurate cpu utilization when sampling at 1 second rate. Otherwise cpu utilization will be only 100 or 0. With this change it can be any value between 0 and 100.

Add srun --accel-bind option to control how tasks are bound to GPUs and NIC
    Generic RESources (GRES).
Based in part upon work by Matthieu Ospici (ATOS).
gres/nic plugin modified to set OMPI_MCA_btl_openib_if_include environment
    variable based upon allocated devices (usable with OpenMPI and Melanox).
Reset GRES env vars after task affinity set

local _get_int functions.

Conflicts:
	src/common/slurm_step_layout.c

afterward cause a divide by zero error.

corruption if thread uses the pointer basing validity off the id.

Bug 1710

Conflicts:
	src/squeue/print.c

If an salloc or srun command is executed on a "front-end" configuration,
    that job will be assigned a slurmd shepherd daemon on the same host as used
    to execute the command when possible rather than an slurmd daemon on an
    arbitrary front-end node.

Disable a test for select/linear and only one node

Disable test with select/linear and only one node

Bug 1495

This implements the same logic changes to the SGI hypercube and
  dragonfly network logic was was done to the regular and topology
  portions of the code in version 14.11, commit
  58623ec7

Conflicts:
	src/sreport/job_reports.c

It was deplicated years ago. The option will still work though.

Correct count of CPUs allocated to job on system with hyperthreads.
The bug was introduced in commit a6d3074d
On a system with hyperthreads:
srun -n1 --ntasks-per-core=1 hostname
you would get:
slurmctld: error: job_update_cpu_cnt: cpu_cnt underflow on job_id 67072

preempt/job_prio plugin: Implement the concept of Warm-up Time here. Use
the QoS GraceTime as the amount of time to wait before preempting.
Basically, skip preemption if your time is not up.

Conflicts:
	src/plugins/accounting_storage/mysql/as_mysql_rollup.c
	src/slurmctld/acct_policy.c

-W/--wait is only supported by srun and should not show
up in the usage string of sbatch or salloc.