GPROF(1) UNIX Programmer's Manual GPROF(1)
gprof - display call graph profile data
S
\bSY
\bYN
\bNO
\bOP
\bPS
\bSI
\bIS
\bS
g
\bgp
\bpr
\bro
\bof
\bf [ options ] [ a.out [ gmon.out ... ] ]
D
\bDE
\bES
\bSC
\bCR
\bRI
\bIP
\bPT
\bTI
\bIO
\bON
\bN
_
\bg_
\bp_
\br_
\bo_
\bf produces an execution profile of C, Pascal, or For-
tran77 programs. The effect of called routines is incor-
porated in the profile of each caller. The profile data is
taken from the call graph profile file (_
\bg_
\bm_
\bo_
\bn._
\bo_
\bu_
\bt default)
which is created by programs that are compiled with the -
\b-p
\bpg
\bg
option of _
\bc_
\bc, _
\bp_
\bc, and _
\bf_
\b7_
\b7. The -
\b-p
\bpg
\bg option also links in
versions of the library routines that are compiled for pro-
filing. The symbol table in the named object file (_
\ba._
\bo_
\bu_
\bt
default) is read and correlated with the call graph profile
file. If more than one profile file is specified, the _
\bg_
\bp_
\br_
\bo_
\bf
output shows the sum of the profile information in the given
_
\bG_
\bp_
\br_
\bo_
\bf 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 descen-
dents. 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 descendents is pro-
pagated 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
_
\bp_
\br_
\bo_
\bf(1). This listing gives the total execution times, the
call counts, the number of milliseconds per call in the rou-
tine itself, and the number of milliseconds per call in the
routine itself including its descendents.
Finally, an index of the function names is provided.
The following options are available:
-
\b-a
\ba suppresses the printing of statically declared func-
tions. If this option is given, all relevant informa-
tion about the static function (_
\be._
\bg., time samples,
calls to other functions, calls from other functions)
belongs to the function loaded just before the static
function in the _
\ba._
\bo_
\bu_
\bt file.
Printed 7/9/88 September 30, 1987 1
GPROF(1) UNIX Programmer's Manual GPROF(1)
-
\b-b
\bb suppresses the printing of a description of each field
-
\b-c
\bc the static call graph of the program is discovered by a
heuristic that examines the text space of the object
file. Static-only parents or children are shown with
-
\b-e
\be _
\bn_
\ba_
\bm_
\be
suppresses the printing of the graph profile entry for
routine _
\bn_
\ba_
\bm_
\be and all its descendants (unless they have
other ancestors that aren't suppressed). More than one
-
\b-e
\be option may be given. Only one _
\bn_
\ba_
\bm_
\be may be given
with each -
\b-e
\be option.
-
\b-E
\bE _
\bn_
\ba_
\bm_
\be
suppresses the printing of the graph profile entry for
routine _
\bn_
\ba_
\bm_
\be (and its descendants) as -
\b-e
\be, above, and
also excludes the time spent in _
\bn_
\ba_
\bm_
\be (and its descen-
dants) from the total and percentage time computations.
(For example, -
\b-E
\bE _
\bm_
\bc_
\bo_
\bu_
\bn_
\bt -
\b-E
\bE _
\bm_
\bc_
\bl_
\be_
\ba_
\bn_
\bu_
\bp is the default.)
-
\b-f
\bf _
\bn_
\ba_
\bm_
\be
prints the graph profile entry of only the specified
routine _
\bn_
\ba_
\bm_
\be and its descendants. More than one -
\b-f
\bf
option may be given. Only one _
\bn_
\ba_
\bm_
\be may be given with
-
\b-F
\bF _
\bn_
\ba_
\bm_
\be
prints the graph profile entry of only the routine _
\bn_
\ba_
\bm_
\be
and its descendants (as -
\b-f
\bf,
\b, above) and also uses only
the times of the printed routines in total time and
percentage computations. More than one -
\b-F
\bF option may
be given. Only one _
\bn_
\ba_
\bm_
\be may be given with each -
\b-F
\bF
option. The -
\b-F
\bF option overrides the -
\b-E
\bE option.
-
\b-k
\bk _
\bf_
\br_
\bo_
\bm_
\bn_
\ba_
\bm_
\be _
\bt_
\bo_
\bn_
\ba_
\bm_
\be
will delete any arcs from routine _
\bf_
\br_
\bo_
\bm_
\bn_
\ba_
\bm_
\be to routine
_
\bt_
\bo_
\bn_
\ba_
\bm_
\be. This can be used to break undesired cycles.
More than one -
\b-k
\bk option may be given. Only one pair of
routine names may be given with each -
\b-k
\bk option.
-
\b-s
\bs a profile file _
\bg_
\bm_
\bo_
\bn._
\bs_
\bu_
\bm is produced that represents the
sum of the profile information in all the specified
profile files. This summary profile file may be given
to later executions of gprof (probably also with a -
\b-s
\bs)
to accumulate profile data across several runs of an
-
\b-z
\bz displays routines that have zero usage (as shown by
call counts and accumulated time). This is useful with
the -
\b-c
\bc option for discovering which routines were never
Printed 7/9/88 September 30, 1987 2
GPROF(1) UNIX Programmer's Manual GPROF(1)
_
\ba._
\bo_
\bu_
\bt the namelist and text space.
_
\bg_
\bm_
\bo_
\bn._
\bo_
\bu_
\bt dynamic call graph and profile.
_
\bg_
\bm_
\bo_
\bn._
\bs_
\bu_
\bm summarized dynamic call graph and profile.
S
\bSE
\bEE
\bE A
\bAL
\bLS
\bSO
\bO
monitor(3), profil(2), cc(1), prof(1)
``gprof: A Call Graph Execution Profiler'', by Graham, S.L.,
Kessler, P.B., McKusick, M.K.; _
\bP_
\br_
\bo_
\bc_
\be_
\be_
\bd_
\bi_
\bn_
\bg_
\bs _
\bo_
\bf _
\bt_
\bh_
\be _
\bS_
\bI_
\bG_
\bP_
\bL_
\bA_
\bN
'_
\b8_
\b2 _
\bS_
\by_
\bm_
\bp_
\bo_
\bs_
\bi_
\bu_
\bm _
\bo_
\bn _
\bC_
\bo_
\bm_
\bp_
\bi_
\bl_
\be_
\br _
\bC_
\bo_
\bn_
\bs_
\bt_
\br_
\bu_
\bc_
\bt_
\bi_
\bo_
\bn, SIGPLAN Notices,
Vol. 17, No. 6, pp. 120-126, June 1982.
Beware of quantization errors. The granularity of the sam-
pling 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 function divided by the
number of times the function is called. Thus the time pro-
pagated along the call graph arcs to the function's parents
is directly proportional to the number of times that arc is
Parents that are not themselves profiled will have the time
of their profiled children propagated to them, but they will
appear to be spontaneously invoked in the call graph list-
ing, and will not have their time propagated further. Simi-
larly, 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
The profiled program must call _
\be_
\bx_
\bi_
\bt(2) or return normally
for the profiling information to be saved in the gmon.out
Printed 7/9/88 September 30, 1987 3
projects / unix-history / blob