----------------------------------------------
lines 6-91 of file: at_cascade/fit_parallel.py
----------------------------------------------

{xrst_begin fit_parallel}
{xrst_spell
  cpus
}

Fit With Specified Maximum Number of Processes
##############################################

Prototype
*********
{xrst_literal
   # BEGIN_DEF
   # END_DEF
}

job_table
*********
This is a :ref:`create_job_table@job_table` containing all the jobs
necessary to fit the :ref:`glossary@fit_goal_set` .

start_job_id
************
This is the :ref:`create_job_table@job_table@job_id`
for the starting job.
The fit_parallel routine will not return until this job,
and all it descendants in the job table, have been run,
or an error occurs that prevents a job from completing..

all_node_database
*****************
:ref:`fit_one_job@all_node_database`

node_table
**********
:ref:`fit_one_job@node_table`

fit_integrand
*************
:ref:`fit_one_job@fit_integrand`

skip_start_job
**************
If this is true (false) the job corresponding to *start_job_id*
will be skipped (will not be skipped).
If this argument is true, the start job must have already been run.
This is useful when using :ref:`continue_cascade-name` .

max_number_cpu
**************
This is the maximum number of cpus (processes) to use.
This must be greater than zero.
If it is one, the jobs are run sequentially; i.e., not in parallel.

fit_type_list
*************
This is a list with one or two elements
and its possible elements are ``both`` and ``fixed``.
For each job, the first type of fit is attempted.
If it fails, and there is a second type of fit, it is attempted.
If it also fails, the corresponding job fails.

shared_unique
*************
All of these jobs us the following two python multiprocessing
shared memory names:

|  *shared_memory_prefix* _ *job_name* *shared_unique* _number_cpu_in_use
|  *shared_memory_prefix* _ *job_name* *shared_unique* _job_status

#. *job_name* is *job_table* [ *start_job_id* ] [ ``"job_name"`` ]

#. :ref:`option_all_table@shared_memory_prefix` is specified
   in the option all table.

#. *shared_unique* is text that makes this shared memory name unique among
   all the currently running calls to fit_parallel.
   It is suggested that you use the empty string for this value unless you
   are running more than one call with the same prefix and job name.

trace.out
*********
If the *max_number_cpu* is one, standard output is not redirected.
Otherwise, standard output for each job is written to a file called
``trace.out`` in the same directory as the database for the job.

{xrst_end fit_parallel}