PROGRAM sum.sis: DEFINES n*VALUE, WHERE n IS TAKEN FROM FIBRE INPUT

COMPILATION:                   osc -maxconcur sum.sis
SINGLE PROCESSOR EXECUTION:    s.out sum.input -r
FOUR PROCESSOR EXECUTION:      s.out sum.input -r -w4

NOTE: osc assumes that n will be 100, so will not concurrentize the
      product loop in sum unless -maxconcur is provided. This is 
      only the case when compiling small kernels like sum.sis and 
      should be used with caution. Since sum.input defines n to be
      40000000, an alternate compilation is:

                    osc -i40000000 sum.sis

      which will yield the appropriate parallelization.
