#  search.condor
#
#  Modified:
#
#    14 October 2005
#
#  Discussion:
#
#    This condor submit description file runs the "search" program.
#
#    The "universe" is "MPI".  This environment is needed in order
#    to run an MPI job.
#
#    The "initial directory" tells CONDOR the name of the working
#    directory.  Although my directory seems to be 
#      /a/fs.csit.fsu.edu/u8/users/burkardt/phoenix_test
#    CONDOR doesn't know about AFS, and I have to specify
#      /home/u8/users/burkardt/phoenix_test
#
#    The "executable" is the program "search", which was created by
#    compiling the FORTRAN77 program "search.f" with mpif77, that is:
#      mpif77 search.f
#      mv a.out search
#    Note that the machine on which the executable runs must be 
#    the same kind of machine on which it was compiled.
#
#    No platform is specified, so the default will be used, that is,
#    the job will be run on a machine which has the same architecture
#    and operating system as the machine from which the job was submitted.
#
#    A log file "search.log" will be produced, containing events that
#    occur during the execution of the job.
#
#    The output command specifies a file into which the output from
#    the program will be written.  $(NODE) is a special macro which
#    will assign a unique value to each machine on which the program
#    is run.  The result is that, probably, we will get four output files,
#    search_0.output, search_1.output, search_2.output and search_3.output.
#
#    The MACHINE_COUNT = 4 command requests that the program be run 
#    on 4 processors.
#
universe = MPI
initialdir = /home/u8/users/burkardt/phoenix_test
executable = search
log = search.log
output = search_$(NODE).output
machine_count = 4
queue