gfortran - Why would non-parallelized Fortran measure too long of CPU time? -

my program showing cpu running time longer time program running, no parallelization written in code.

the code written in fortran 90 (there's 1 or 2 later-fortran things added in) , compiled linux machine's native gfortran compiler (--version information: gnu fortran (gcc) 4.4.7 20120313 (red hat 4.4.7-17) ). understand gfortran compiles later standards 90.

when program starts calls call cpu_time(time_start) , before ended calls call cpu_time(time_end). in case, time_end - time_start gives elapsed cpu time in seconds.

so here's curious thing: used htcondor submit code run on whatever machine in local network had available cpu. htcondor log files show job submitted 07/24 14:17:46, started running 15 seconds later, ran completion on same machine, ending 07/30 11:01:52, clock time of less 6 days. however, time_end - time_start says cpu time 993535 seconds, or on 11 days. code not parallelized @ not understand how can be. how can be?

i've ran code hundreds of times before , never noticed phenomenon, i've never checked closely either.

edit: wish note once again code not parallelized, @ least not explicitly. compiled -o3 flag, don't think introduces parallelization. if linked question/answer parallel fortran indeed answer question serial process, please me understand how because not see connection.

my htcondor submission script follows. condor_submit script , that's how run code.

executable     = /path/to/executable universe       = standard log            = condorlog.log output         = condorstdout.out error          = condorerror.out should_transfer_files = if_needed when_to_transfer_output = on_exit queue