Another program profiling tool:

The idea is that you compile with the -p flag 
(and also use the -p flag on the load line)
It's a good idea to shut off all I/O if possible

When you run the program compiled with the -p option,
 you get a file called mon.out

You then run the "prof" program with the -C argument

	prof -C ExecutableName > output

You must give the name of the executable and you must
also specify the -C flag.  The output looks as follows:

Where %time is the % of total time spent in the routine,
cumsecs gives the amount of time spent on a routine and
all routines above it.  The # calls is self explanatory
ms/call gives the average time per call in miliseconds.

 %time  cumsecs  #call  ms/call  name
  91.0     1.52   1005     1.51  Walk(POINT*, int, int&, int&)
                                 [___0FEWalkP6FPOINTiRiTD]

   4.2     1.59    400     0.17  sbrk
                                 [_sbrk]

   1.2     1.61                  mrand48
                                 [_mrand48]

   0.6     1.62  14071     0.00  .mul
   0.6     1.63      1    10.00  ReportBounds(POINT*, int, int&, int&, int&, int&)
                                 [___0FMReportBoundsP6FPOINTiRiNDD]

   0.6     1.64   2019     0.00  malloc
                                 [_malloc]

   0.6     1.65    673     0.01  strlen
                                 [_strlen]



Using -p can slow the program down, so you shouldn't use this option
unless you really want to get profile information.