As it was described in the How to run multi-core jobs in production
document, we are introducing two command-line options for the job transform:
	--multithreaded
	--multiprocess

As the names suggest, the former is used for running AthenaMT and the latter
for running AthenaMP. The number of threads/processes is taken from a new
ATHENA_CORE_NUMBER environment.

Now, we have two equivalent ways to run a job in AthenaMT
(Reco_tf is used here as an example):

ATHENA_CORE_NUMBER=N Reco_tf.py --multithreaded [Other Options]
Reco_tf.py --athenaopts=--threads=N [Other Options]

Similarly, for running AthenaMP we have two options:

ATHENA_CORE_NUMBER=N Reco_tf.py	--multiprocess [Other Options]
Reco_tf.py --athenaopts=--nprocs=N [Other Options]

We will get a configuration error from the transform if one of the following occurs:

* ATHENA_PROC_NUMBER was not set but either --multiprocess or --multithreaded was provided
* Both --multiprocess and --multithreaded options were provided to the same job
* --multiprocess was used in combination with either --athenaopts=--nprocs=N or --athenaopts=--threads=N
* --multithreaded was used in combination with --athenaopts=--threads=N or --athenaopts=--nprocs=N

By implementing	these changes we declare the old ATHENA_RPOC_NUMBER environment
obsolete. Athena repository has	been cleaned up	from the dependency on this
environment. The only exception is Generator code, which is yet to be migrated.