GPROF(1) BSD General Commands Manual GPROF(1)
gprof -- display call graph profile data
gprof [-abKlLsuz] [-C count] [-e name] [-E name] [-f name] [-F name] [-k fromname toname]
[a.out [a.out.gmon ...]]
The gprof utility produces an execution profile of C, Pascal, or Fortran77 programs. The
effect of called routines is incorporated in the profile of each caller. The profile data
is taken from the call graph profile file which is created by programs that are compiled
with the -pg option of cc(1), pc(1), and f77(1). The -pg option also links in versions of
the library routines that are compiled for profiling. By convention these libraries have
their name suffixed with _p, i.e., the profiled version of libc.a is libc_p.a and if you
specify libraries directly to the compiler or linker you can use -lc_p instead of -lc. Read
the given object file (the default is a.out) and establishes the relation between its symbol
table and the call graph profile. The default graph profile file name is the name of the
executable with the suffix .gmon appended. If more than one profile file is specified, the
gprof output shows the sum of the profile information in the given profile files.
The gprof utility calculates the amount of time spent in each routine. Next, these times
are propagated along the edges of the call graph. Cycles are discovered, and calls into a
cycle are made to share the time of the cycle. The first listing shows the functions sorted
according to the time they represent including the time of their call graph descendants.
Below each function entry is shown its (direct) call graph children, and how their times are
propagated to this function. A similar display above the function shows how this function's
time and the time of its descendants is propagated to its (direct) call graph parents.
Cycles are also shown, with an entry for the cycle as a whole and a listing of the members
of the cycle and their contributions to the time and call counts of the cycle.
Second, a flat profile is given, similar to that provided by prof(1). This listing gives
the total execution times, the call counts, the time that the call spent in the routine
itself, and the time that the call spent in the routine itself including its descendants.
The units for the per-call times are normally milliseconds, but they are nanoseconds if the
profiling clock frequency is 10 million or larger, and if a function appears to be never
called then its total self time is printed as a percentage in the self time per call column.
The very high profiling clock frequencies needed to get sufficient accuracy in the per-call
times for short-lived programs are only implemented for ``high resolution'' (non-statisti-
cal) kernel profiling.
Finally, an index of the function names is provided.
The following options are available:
-a Suppress the printing of statically declared functions. If this option is given,
all relevant information about the static function (e.g., time samples, calls to
other functions, calls from other functions) belongs to the function loaded just
before the static function in the a.out file.
-b Suppress the printing of a description of each field in the profile.
Find a minimal set of arcs that can be broken to eliminate all cycles with count or
more members. Caution: the algorithm used to break cycles is exponential, so using
this option may cause gprof to run for a very long time.
Suppress the printing of the graph profile entry for routine name and all its
descendants (unless they have other ancestors that are not suppressed). More than
one -e option may be given. Only one name may be given with each -e option.
Suppress the printing of the graph profile entry for routine name (and its descen-
dants) as -e, above, and also excludes the time spent in name (and its descendants)
from the total and percentage time computations. (For example, -E mcount -E
mcleanup is the default.)
Print the graph profile entry of only the specified routine name and its descen-
dants. More than one -f option may be given. Only one name may be given with each
Print the graph profile entry of only the routine name and its descendants (as -f,
above) and also uses only the times of the printed routines in total time and per-
centage computations. More than one -F option may be given. Only one name may be
given with each -F option. The -F option overrides the -E option.
-k fromname toname
Will delete any arcs from routine fromname to routine toname. This can be used to
break undesired cycles. More than one -k option may be given. Only one pair of
routine names may be given with each -k option.
-K Gather information about symbols from the currently-running kernel using the
sysctl(3) and kldsym(2) interfaces. This forces the a.out argument to be ignored,
and allows for symbols in kld(4) modules to be used.
-l Suppress the printing of the call-graph profile.
-L Suppress the printing of the flat profile.
-s A profile file gmon.sum is produced that represents the sum of the profile informa-
tion in all the specified profile files. This summary profile file may be given to
later executions of gprof (probably also with a -s) to accumulate profile data
across several runs of an a.out file.
-u Suppress the printing of functions whose names are not visible to C programs. For
the ELF object format, this means names that contain the '.' character. For the
a.out object format, it means names that do not begin with a '_' character. All
relevant information about such functions belongs to the (non-suppressed) function
with the next lowest address. This is useful for eliminating "functions" that are
just labels inside other functions.
-z Display routines that have zero usage (as shown by call counts and accumulated
a.out The namelist and text space.
a.out.gmon Dynamic call graph and profile.
gmon.sum Summarized dynamic call graph and profile.
cc(1), profil(2), clocks(7)
S. Graham, P. Kessler, and M. McKusick, "An Execution Profiler for Modular Programs",
Software - Practice and Experience, 13, pp. 671-685, 1983.
S. Graham, P. Kessler, and M. McKusick, "gprof: A Call Graph Execution Profiler",
Proceedings of the SIGPLAN '82 Symposium on Compiler Construction, SIGPLAN Notices, 6, 17,
pp. 120-126, June 1982.
The gprof profiler appeared in 4.2BSD.
The granularity of the sampling is shown, but remains statistical at best. We assume that
the time for each execution of a function can be expressed by the total time for the func-
tion divided by the number of times the function is called. Thus the time propagated along
the call graph arcs to the function's parents is directly proportional to the number of
times that arc is traversed.
Parents that are not themselves profiled will have the time of their profiled children prop-
agated to them, but they will appear to be spontaneously invoked in the call graph listing,
and will not have their time propagated further. Similarly, signal catchers, even though
profiled, will appear to be spontaneous (although for more obscure reasons). Any profiled
children of signal catchers should have their times propagated properly, unless the signal
catcher was invoked during the execution of the profiling routine, in which case all is
The profiled program must call exit(3) or return normally for the profiling information to
be saved in the graph profile file.
BSD December 25, 2008 BSD