RT-Thread-Studio
/
sdk-toolchain-GNU-ARM-Embedded
镜像来自 https://github-proxy.rt-thread.io/RT-Thread-Studio/sdk-toolchain-GNU-ARM-Embedded.git


			
							12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788899091929394959697989910010110210310410510610710810911011111211311411511611711811912012112212312412512612712812913013113213313413513613713813914014114214314414514614714814915015115215315415515615715815916016116216316416516616716816917017117217317417517617717817918018118218318418518618718818919019119219319419519619719819920020120220320420520620720820921021121221321421521621721821922022122222322422522622722822923023123223323423523623723823924024124224324424524624724824925025125225325425525625725825926026126226326426526626726826927027127227327427527627727827928028128228328428528628728828929029129229329429529629729829930030130230330430530630730830931031131231331431531631731831932032132232332432532632732832933033133233333433533633733833934034134234334434534634734834935035135235335435535635735835936036136236336436536636736836937037137237337437537637737837938038138238338438538638738838939039139239339439539639739839940040140240340440540640740840941041141241341441541641741841942042142242342442542642742842943043143243343443543643743843944044144244344444544644744844945045145245345445545645745845946046146246346446546646746846947047147247347447547647747847948048148248348448548648748848949049149249349449549649749849950050150250350450550650750850951051151251351451551651751851952052152252352452552652752852953053153253353453553653753853954054154254354454554654754854955055155255355455555655755855956056156256356456556656756856957057157257357457557657757857958058158258358458558658758858959059159259359459559659759859960060160260360460560660760860961061161261361461561661761861962062162262362462562662762862963063163263363463563663763863964064164264364464564664764864965065165265365465565665765865966066166266366466566666766866967067167267367467567667767867968068168268368468568668768868969069169269369469569669769869970070170270370470570670770870971071171271371471571671771871972072172272372472572672772872973073173273373473573673773873974074174274374474574674774874975075175275375475575675775875976076176276376476576676776876977077177277377477577677777877978078178278378478578678778878979079179279379479579679779879980080180280380480580680780880981081181281381481581681781881982082182282382482582682782882983083183283383483583683783883984084184284384484584684784884985085185285385485585685785885986086186286386486586686786886987087187287387487587687787887988088188288388488588688788888989089189289389489589689789889990090190290390490590690790890991091191291391491591691791891992092192292392492592692792892993093193293393493593693793893994094194294394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001100210031004100510061007100810091010101110121013101410151016101710181019102010211022102310241025102610271028102910301031103210331034103510361037103810391040104110421043104410451046104710481049105010511052105310541055105610571058105910601061106210631064106510661067106810691070107110721073107410751076107710781079108010811082108310841085108610871088108910901091109210931094109510961097109810991100110111021103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175117611771178117911801181118211831184118511861187118811891190119111921193119411951196119711981199120012011202120312041205120612071208120912101211121212131214121512161217121812191220122112221223122412251226122712281229123012311232123312341235123612371238123912401241124212431244124512461247124812491250125112521253125412551256125712581259126012611262126312641265126612671268126912701271127212731274127512761277127812791280128112821283128412851286128712881289129012911292129312941295129612971298129913001301130213031304130513061307130813091310131113121313131413151316131713181319132013211322132313241325132613271328132913301331133213331334133513361337133813391340134113421343134413451346134713481349135013511352135313541355135613571358135913601361136213631364136513661367136813691370137113721373137413751376137713781379138013811382138313841385138613871388138913901391139213931394139513961397139813991400140114021403140414051406140714081409141014111412141314141415141614171418141914201421142214231424142514261427142814291430143114321433143414351436143714381439144014411442144314441445144614471448144914501451145214531454145514561457145814591460146114621463146414651466146714681469147014711472147314741475147614771478147914801481148214831484148514861487148814891490149114921493149414951496149714981499150015011502150315041505150615071508150915101511151215131514151515161517151815191520152115221523152415251526152715281529153015311532153315341535153615371538153915401541154215431544154515461547154815491550155115521553155415551556155715581559156015611562156315641565156615671568156915701571157215731574157515761577157815791580158115821583158415851586158715881589159015911592159315941595159615971598159916001601160216031604160516061607160816091610161116121613161416151616161716181619162016211622162316241625162616271628162916301631163216331634163516361637163816391640164116421643164416451646164716481649165016511652165316541655165616571658165916601661166216631664166516661667166816691670167116721673167416751676167716781679168016811682168316841685168616871688168916901691169216931694169516961697169816991700170117021703170417051706170717081709171017111712171317141715171617171718171917201721172217231724172517261727172817291730173117321733173417351736173717381739174017411742174317441745174617471748174917501751175217531754175517561757175817591760176117621763176417651766176717681769177017711772177317741775177617771778177917801781178217831784178517861787178817891790179117921793179417951796179717981799180018011802180318041805180618071808180918101811181218131814181518161817181818191820182118221823182418251826182718281829183018311832183318341835183618371838183918401841184218431844184518461847184818491850185118521853185418551856185718581859186018611862186318641865186618671868186918701871187218731874187518761877187818791880188118821883188418851886188718881889189018911892189318941895189618971898189919001901190219031904190519061907190819091910191119121913191419151916191719181919192019211922192319241925192619271928192919301931193219331934193519361937193819391940194119421943194419451946194719481949195019511952195319541955195619571958195919601961196219631964196519661967196819691970197119721973197419751976197719781979198019811982198319841985198619871988198919901991199219931994199519961997199819992000200120022003200420052006200720082009201020112012201320142015201620172018201920202021202220232024202520262027202820292030203120322033203420352036203720382039204020412042204320442045204620472048204920502051205220532054205520562057205820592060206120622063206420652066206720682069207020712072207320742075207620772078207920802081208220832084208520862087208820892090209120922093209420952096209720982099210021012102210321042105210621072108210921102111211221132114211521162117211821192120212121222123212421252126212721282129213021312132213321342135213621372138213921402141214221432144214521462147214821492150215121522153215421552156215721582159216021612162216321642165216621672168216921702171217221732174217521762177217821792180218121822183218421852186218721882189219021912192219321942195219621972198219922002201220222032204220522062207220822092210221122122213221422152216221722182219222022212222222322242225222622272228222922302231223222332234223522362237223822392240224122422243224422452246224722482249225022512252225322542255225622572258225922602261226222632264226522662267226822692270227122722273227422752276227722782279228022812282228322842285228622872288228922902291229222932294229522962297229822992300230123022303230423052306230723082309231023112312231323142315231623172318231923202321232223232324232523262327232823292330233123322333233423352336233723382339234023412342234323442345234623472348234923502351235223532354235523562357235823592360236123622363236423652366236723682369237023712372237323742375237623772378237923802381238223832384238523862387238823892390239123922393239423952396239723982399240024012402240324042405240624072408240924102411241224132414241524162417241824192420242124222423242424252426242724282429243024312432243324342435243624372438243924402441244224432444244524462447244824492450245124522453245424552456245724582459246024612462246324642465246624672468246924702471247224732474247524762477247824792480248124822483248424852486248724882489249024912492249324942495249624972498249925002501250225032504250525062507250825092510251125122513251425152516251725182519252025212522252325242525252625272528252925302531253225332534253525362537253825392540254125422543254425452546254725482549255025512552255325542555255625572558255925602561256225632564256525662567256825692570257125722573257425752576257725782579258025812582258325842585258625872588258925902591259225932594259525962597259825992600260126022603260426052606260726082609261026112612261326142615261626172618261926202621262226232624262526262627262826292630263126322633263426352636263726382639264026412642264326442645264626472648264926502651265226532654265526562657265826592660266126622663266426652666266726682669267026712672267326742675267626772678267926802681268226832684268526862687268826892690269126922693269426952696269726982699270027012702270327042705270627072708270927102711271227132714271527162717271827192720272127222723272427252726272727282729273027312732273327342735273627372738273927402741274227432744274527462747274827492750275127522753275427552756275727582759276027612762276327642765276627672768276927702771277227732774277527762777277827792780278127822783278427852786278727882789279027912792279327942795279627972798279928002801280228032804280528062807280828092810281128122813281428152816281728182819282028212822282328242825282628272828282928302831283228332834
							<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<!-- This file documents the gprof profiler of the GNU system.

Copyright (C) 1988-2024 Free Software Foundation, Inc.

Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3
or any later version published by the Free Software Foundation;
with no Invariant Sections, with no Front-Cover Texts, and with no
Back-Cover Texts.  A copy of the license is included in the
section entitled "GNU Free Documentation License".
 -->
<!-- Created by GNU Texinfo 6.7, http://www.gnu.org/software/texinfo/ -->
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<title>GNU gprof</title>

<meta name="description" content="GNU gprof">
<meta name="keywords" content="GNU gprof">
<meta name="resource-type" content="document">
<meta name="distribution" content="global">
<meta name="Generator" content="makeinfo">
<link href="#Top" rel="start" title="Top">
<link href="#SEC_Contents" rel="contents" title="Table of Contents">
<style type="text/css">
<!--
a.summary-letter {text-decoration: none}
blockquote.indentedblock {margin-right: 0em}
div.display {margin-left: 3.2em}
div.example {margin-left: 3.2em}
div.lisp {margin-left: 3.2em}
kbd {font-style: oblique}
pre.display {font-family: inherit}
pre.format {font-family: inherit}
pre.menu-comment {font-family: serif}
pre.menu-preformatted {font-family: serif}
span.nolinebreak {white-space: nowrap}
span.roman {font-family: initial; font-weight: normal}
span.sansserif {font-family: sans-serif; font-weight: normal}
ul.no-bullet {list-style: none}
-->
</style>


</head>

<body lang="en">
<h1 class="settitle" align="center">GNU gprof</h1>


<span id="SEC_Contents"></span>
<h2 class="contents-heading">Table of Contents</h2>

<div class="contents">

<ul class="no-bullet">
  <li><a id="toc-Introduction-to-Profiling" href="#Introduction">1 Introduction to Profiling</a></li>
  <li><a id="toc-Compiling-a-Program-for-Profiling" href="#Compiling">2 Compiling a Program for Profiling</a></li>
  <li><a id="toc-Executing-the-Program" href="#Executing">3 Executing the Program</a></li>
  <li><a id="toc-gprof-Command-Summary" href="#Invoking">4 <code>gprof</code> Command Summary</a>
  <ul class="no-bullet">
    <li><a id="toc-Output-Options-1" href="#Output-Options">4.1 Output Options</a></li>
    <li><a id="toc-Analysis-Options-1" href="#Analysis-Options">4.2 Analysis Options</a></li>
    <li><a id="toc-Miscellaneous-Options-1" href="#Miscellaneous-Options">4.3 Miscellaneous Options</a></li>
    <li><a id="toc-Deprecated-Options-1" href="#Deprecated-Options">4.4 Deprecated Options</a></li>
    <li><a id="toc-Symspecs-1" href="#Symspecs">4.5 Symspecs</a></li>
  </ul></li>
  <li><a id="toc-Interpreting-gprof_0027s-Output" href="#Output">5 Interpreting <code>gprof</code>&rsquo;s Output</a>
  <ul class="no-bullet">
    <li><a id="toc-The-Flat-Profile" href="#Flat-Profile">5.1 The Flat Profile</a></li>
    <li><a id="toc-The-Call-Graph" href="#Call-Graph">5.2 The Call Graph</a>
    <ul class="no-bullet">
      <li><a id="toc-The-Primary-Line" href="#Primary">5.2.1 The Primary Line</a></li>
      <li><a id="toc-Lines-for-a-Function_0027s-Callers" href="#Callers">5.2.2 Lines for a Function&rsquo;s Callers</a></li>
      <li><a id="toc-Lines-for-a-Function_0027s-Subroutines" href="#Subroutines">5.2.3 Lines for a Function&rsquo;s Subroutines</a></li>
      <li><a id="toc-How-Mutually-Recursive-Functions-Are-Described" href="#Cycles">5.2.4 How Mutually Recursive Functions Are Described</a></li>
    </ul></li>
    <li><a id="toc-Line_002dby_002dline-Profiling" href="#Line_002dby_002dline">5.3 Line-by-line Profiling</a></li>
    <li><a id="toc-The-Annotated-Source-Listing" href="#Annotated-Source">5.4 The Annotated Source Listing</a></li>
  </ul></li>
  <li><a id="toc-Inaccuracy-of-gprof-Output" href="#Inaccuracy">6 Inaccuracy of <code>gprof</code> Output</a>
  <ul class="no-bullet">
    <li><a id="toc-Statistical-Sampling-Error" href="#Sampling-Error">6.1 Statistical Sampling Error</a></li>
    <li><a id="toc-Estimating-children-Times" href="#Assumptions">6.2 Estimating <code>children</code> Times</a></li>
  </ul></li>
  <li><a id="toc-Answers-to-Common-Questions" href="#How-do-I_003f">7 Answers to Common Questions</a></li>
  <li><a id="toc-Incompatibilities-with-Unix-gprof" href="#Incompatibilities">8 Incompatibilities with Unix <code>gprof</code></a></li>
  <li><a id="toc-Details-of-Profiling" href="#Details">9 Details of Profiling</a>
  <ul class="no-bullet">
    <li><a id="toc-Implementation-of-Profiling" href="#Implementation">9.1 Implementation of Profiling</a></li>
    <li><a id="toc-Profiling-Data-File-Format" href="#File-Format">9.2 Profiling Data File Format</a>
    <ul class="no-bullet">
      <li><a id="toc-Histogram-Records" href="#Histogram-Records">9.2.1 Histogram Records</a></li>
      <li><a id="toc-Call_002dGraph-Records" href="#Call_002dGraph-Records">9.2.2 Call-Graph Records</a></li>
      <li><a id="toc-Basic_002dBlock-Execution-Count-Records" href="#Basic_002dBlock-Execution-Count-Records">9.2.3 Basic-Block Execution Count Records</a></li>
    </ul></li>
    <li><a id="toc-gprof_0027s-Internal-Operation" href="#Internals">9.3 <code>gprof</code>&rsquo;s Internal Operation</a></li>
    <li><a id="toc-Debugging-gprof" href="#Debugging">9.4 Debugging <code>gprof</code></a></li>
  </ul></li>
  <li><a id="toc-GNU-Free-Documentation-License-1" href="#GNU-Free-Documentation-License">Appendix A GNU Free Documentation License</a></li>
</ul>
</div>


<span id="Top"></span><div class="header">
<p>
Next: <a href="#Introduction" accesskey="n" rel="next">Introduction</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Profiling-a-Program_003a-Where-Does-It-Spend-Its-Time_003f"></span><h1 class="top">Profiling a Program: Where Does It Spend Its Time?</h1>

<p>This manual describes the <small>GNU</small> profiler, <code>gprof</code>, and how you
can use it to determine which parts of a program are taking most of the
execution time.  We assume that you know how to write, compile, and
execute programs.  <small>GNU</small> <code>gprof</code> was written by Jay Fenlason.
</p>
<p>This manual is for <code>gprof</code>
(Arm GNU Toolchain 13.3.Rel1 (Build arm-13.24))
version 2.42.0.
</p>
<p>This document is distributed under the terms of the GNU Free
Documentation License version 1.3.  A copy of the license is included
in the section entitled &ldquo;GNU Free Documentation License&rdquo;.
</p>
<table class="menu" border="0" cellspacing="0">
<tr><td align="left" valign="top">&bull; <a href="#Introduction" accesskey="1">Introduction</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">What profiling means, and why it is useful.
</td></tr>
<tr><th colspan="3" align="left" valign="top"><pre class="menu-comment">

</pre></th></tr><tr><td align="left" valign="top">&bull; <a href="#Compiling" accesskey="2">Compiling</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">How to compile your program for profiling.
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Executing" accesskey="3">Executing</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Executing your program to generate profile data
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Invoking" accesskey="4">Invoking</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">How to run <code>gprof</code>, and its options
</td></tr>
<tr><th colspan="3" align="left" valign="top"><pre class="menu-comment">

</pre></th></tr><tr><td align="left" valign="top">&bull; <a href="#Output" accesskey="5">Output</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Interpreting <code>gprof</code>&rsquo;s output
</td></tr>
<tr><th colspan="3" align="left" valign="top"><pre class="menu-comment">

</pre></th></tr><tr><td align="left" valign="top">&bull; <a href="#Inaccuracy" accesskey="6">Inaccuracy</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Potential problems you should be aware of
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#How-do-I_003f" accesskey="7">How do I?</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Answers to common questions
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Incompatibilities" accesskey="8">Incompatibilities</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">(between <small>GNU</small> <code>gprof</code> and Unix <code>gprof</code>.)
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Details" accesskey="9">Details</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Details of how profiling is done
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#GNU-Free-Documentation-License">GNU Free Documentation License</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">GNU Free Documentation License
</td></tr>
</table>

<hr>
<span id="Introduction"></span><div class="header">
<p>
Next: <a href="#Compiling" accesskey="n" rel="next">Compiling</a>, Previous: <a href="#Top" accesskey="p" rel="prev">Top</a>, Up: <a href="#Top" accesskey="u" rel="up">Top</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Introduction-to-Profiling"></span><h2 class="chapter">1 Introduction to Profiling</h2>


<p>Profiling allows you to learn where your program spent its time and which
functions called which other functions while it was executing.  This
information can show you which pieces of your program are slower than you
expected, and might be candidates for rewriting to make your program
execute faster.  It can also tell you which functions are being called more
or less often than you expected.  This may help you spot bugs that had
otherwise been unnoticed.
</p>
<p>Since the profiler uses information collected during the actual execution
of your program, it can be used on programs that are too large or too
complex to analyze by reading the source.  However, how your program is run
will affect the information that shows up in the profile data.  If you
don&rsquo;t use some feature of your program while it is being profiled, no
profile information will be generated for that feature.
</p>
<p>Profiling has several steps:
</p>
<ul>
<li> You must compile and link your program with profiling enabled.
See <a href="#Compiling">Compiling a Program for Profiling</a>.

</li><li> You must execute your program to generate a profile data file.
See <a href="#Executing">Executing the Program</a>.

</li><li> You must run <code>gprof</code> to analyze the profile data.
See <a href="#Invoking"><code>gprof</code> Command Summary</a>.
</li></ul>

<p>The next three chapters explain these steps in greater detail.
</p>

<p>Several forms of output are available from the analysis.
</p>
<p>The <em>flat profile</em> shows how much time your program spent in each function,
and how many times that function was called.  If you simply want to know
which functions burn most of the cycles, it is stated concisely here.
See <a href="#Flat-Profile">The Flat Profile</a>.
</p>
<p>The <em>call graph</em> shows, for each function, which functions called it, which
other functions it called, and how many times.  There is also an estimate
of how much time was spent in the subroutines of each function.  This can
suggest places where you might try to eliminate function calls that use a
lot of time.  See <a href="#Call-Graph">The Call Graph</a>.
</p>
<p>The <em>annotated source</em> listing is a copy of the program&rsquo;s
source code, labeled with the number of times each line of the
program was executed.  See <a href="#Annotated-Source">The Annotated Source
Listing</a>.
</p>
<p>To better understand how profiling works, you may wish to read
a description of its implementation.
See <a href="#Implementation">Implementation of Profiling</a>.
</p>
<hr>
<span id="Compiling"></span><div class="header">
<p>
Next: <a href="#Executing" accesskey="n" rel="next">Executing</a>, Previous: <a href="#Introduction" accesskey="p" rel="prev">Introduction</a>, Up: <a href="#Top" accesskey="u" rel="up">Top</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Compiling-a-Program-for-Profiling"></span><h2 class="chapter">2 Compiling a Program for Profiling</h2>

<p>The first step in generating profile information for your program is
to compile and link it with profiling enabled.
</p>
<p>To compile a source file for profiling, specify the &lsquo;<samp>-pg</samp>&rsquo; option when
you run the compiler.  (This is in addition to the options you normally
use.)
</p>
<p>To link the program for profiling, if you use a compiler such as <code>cc</code>
to do the linking, simply specify &lsquo;<samp>-pg</samp>&rsquo; in addition to your usual
options.  The same option, &lsquo;<samp>-pg</samp>&rsquo;, alters either compilation or linking
to do what is necessary for profiling.  Here are examples:
</p>
<div class="example">
<pre class="example">cc -g -c myprog.c utils.c -pg
cc -o myprog myprog.o utils.o -pg
</pre></div>

<p>The &lsquo;<samp>-pg</samp>&rsquo; option also works with a command that both compiles and links:
</p>
<div class="example">
<pre class="example">cc -o myprog myprog.c utils.c -g -pg
</pre></div>

<p>Note: The &lsquo;<samp>-pg</samp>&rsquo; option must be part of your compilation options
as well as your link options.  If it is not then no call-graph data
will be gathered and when you run <code>gprof</code> you will get an error
message like this:
</p>
<div class="example">
<pre class="example">gprof: gmon.out file is missing call-graph data
</pre></div>

<p>If you add the &lsquo;<samp>-Q</samp>&rsquo; switch to suppress the printing of the call
graph data you will still be able to see the time samples:
</p>
<div class="example">
<pre class="example">Flat profile:

Each sample counts as 0.01 seconds.
  %   cumulative   self              self     total
 time   seconds   seconds    calls  Ts/call  Ts/call  name
 44.12      0.07     0.07                             zazLoop
 35.29      0.14     0.06                             main
 20.59      0.17     0.04                             bazMillion
</pre></div>

<p>If you run the linker <code>ld</code> directly instead of through a compiler
such as <code>cc</code>, you may have to specify a profiling startup file
<samp>gcrt0.o</samp> as the first input file instead of the usual startup
file <samp>crt0.o</samp>.  In addition, you would probably want to
specify the profiling C library, <samp>libc_p.a</samp>, by writing
&lsquo;<samp>-lc_p</samp>&rsquo; instead of the usual &lsquo;<samp>-lc</samp>&rsquo;.  This is not absolutely
necessary, but doing this gives you number-of-calls information for
standard library functions such as <code>read</code> and <code>open</code>.  For
example:
</p>
<div class="example">
<pre class="example">ld -o myprog /lib/gcrt0.o myprog.o utils.o -lc_p
</pre></div>

<p>If you are running the program on a system which supports shared
libraries you may run into problems with the profiling support code in
a shared library being called before that library has been fully
initialised.  This is usually detected by the program encountering a
segmentation fault as soon as it is run.  The solution is to link
against a static version of the library containing the profiling
support code, which for <code>gcc</code> users can be done via the
&lsquo;<samp>-static</samp>&rsquo; or &lsquo;<samp>-static-libgcc</samp>&rsquo; command-line option.  For
example:
</p>
<div class="example">
<pre class="example">gcc -g -pg -static-libgcc myprog.c utils.c -o myprog
</pre></div>

<p>If you compile only some of the modules of the program with &lsquo;<samp>-pg</samp>&rsquo;, you
can still profile the program, but you won&rsquo;t get complete information about
the modules that were compiled without &lsquo;<samp>-pg</samp>&rsquo;.  The only information
you get for the functions in those modules is the total time spent in them;
there is no record of how many times they were called, or from where.  This
will not affect the flat profile (except that the <code>calls</code> field for
the functions will be blank), but will greatly reduce the usefulness of the
call graph.
</p>
<p>If you wish to perform line-by-line profiling you should use the
<code>gcov</code> tool instead of <code>gprof</code>.  See that tool&rsquo;s manual or
info pages for more details of how to do this.
</p>
<p>Note, older versions of <code>gcc</code> produce line-by-line profiling
information that works with <code>gprof</code> rather than <code>gcov</code> so
there is still support for displaying this kind of information in
<code>gprof</code>. See <a href="#Line_002dby_002dline">Line-by-line Profiling</a>.
</p>
<p>It also worth noting that <code>gcc</code> implements a
&lsquo;<samp>-finstrument-functions</samp>&rsquo; command-line option which will insert
calls to special user supplied instrumentation routines at the entry
and exit of every function in their program.  This can be used to
implement an alternative profiling scheme.
</p>
<hr>
<span id="Executing"></span><div class="header">
<p>
Next: <a href="#Invoking" accesskey="n" rel="next">Invoking</a>, Previous: <a href="#Compiling" accesskey="p" rel="prev">Compiling</a>, Up: <a href="#Top" accesskey="u" rel="up">Top</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Executing-the-Program"></span><h2 class="chapter">3 Executing the Program</h2>

<p>Once the program is compiled for profiling, you must run it in order to
generate the information that <code>gprof</code> needs.  Simply run the program
as usual, using the normal arguments, file names, etc.  The program should
run normally, producing the same output as usual.  It will, however, run
somewhat slower than normal because of the time spent collecting and
writing the profile data.
</p>
<p>The way you run the program&mdash;the arguments and input that you give
it&mdash;may have a dramatic effect on what the profile information shows.  The
profile data will describe the parts of the program that were activated for
the particular input you use.  For example, if the first command you give
to your program is to quit, the profile data will show the time used in
initialization and in cleanup, but not much else.
</p>
<p>Your program will write the profile data into a file called <samp>gmon.out</samp>
just before exiting.  If there is already a file called <samp>gmon.out</samp>,
its contents are overwritten.  You can rename the file afterwards if you
are concerned that it may be overwritten.  If your system libc allows you
may be able to write the profile data under a different name.  Set the
GMON_OUT_PREFIX environment variable; this name will be appended with
the PID of the running program.
</p>
<p>In order to write the <samp>gmon.out</samp> file properly, your program must exit
normally: by returning from <code>main</code> or by calling <code>exit</code>.  Calling
the low-level function <code>_exit</code> does not write the profile data, and
neither does abnormal termination due to an unhandled signal.
</p>
<p>The <samp>gmon.out</samp> file is written in the program&rsquo;s <em>current working
directory</em> at the time it exits.  This means that if your program calls
<code>chdir</code>, the <samp>gmon.out</samp> file will be left in the last directory
your program <code>chdir</code>&rsquo;d to.  If you don&rsquo;t have permission to write in
this directory, the file is not written, and you will get an error message.
</p>
<p>Older versions of the <small>GNU</small> profiling library may also write a file
called <samp>bb.out</samp>.  This file, if present, contains an human-readable
listing of the basic-block execution counts.  Unfortunately, the
appearance of a human-readable <samp>bb.out</samp> means the basic-block
counts didn&rsquo;t get written into <samp>gmon.out</samp>.
The Perl script <code>bbconv.pl</code>, included with the <code>gprof</code>
source distribution, will convert a <samp>bb.out</samp> file into
a format readable by <code>gprof</code>.  Invoke it like this:
</p>
<div class="example">
<pre class="example">bbconv.pl &lt; bb.out &gt; <var>bh-data</var>
</pre></div>

<p>This translates the information in <samp>bb.out</samp> into a form that
<code>gprof</code> can understand.  But you still need to tell <code>gprof</code>
about the existence of this translated information.  To do that, include
<var>bb-data</var> on the <code>gprof</code> command line, <em>along with
<samp>gmon.out</samp></em>, like this:
</p>
<div class="example">
<pre class="example">gprof <var>options</var> <var>executable-file</var> gmon.out <var>bb-data</var> [<var>yet-more-profile-data-files</var>&hellip;] [&gt; <var>outfile</var>]
</pre></div>

<hr>
<span id="Invoking"></span><div class="header">
<p>
Next: <a href="#Output" accesskey="n" rel="next">Output</a>, Previous: <a href="#Executing" accesskey="p" rel="prev">Executing</a>, Up: <a href="#Top" accesskey="u" rel="up">Top</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="gprof-Command-Summary"></span><h2 class="chapter">4 <code>gprof</code> Command Summary</h2>

<p>After you have a profile data file <samp>gmon.out</samp>, you can run <code>gprof</code>
to interpret the information in it.  The <code>gprof</code> program prints a
flat profile and a call graph on standard output.  Typically you would
redirect the output of <code>gprof</code> into a file with &lsquo;<samp>&gt;</samp>&rsquo;.
</p>
<p>You run <code>gprof</code> like this:
</p>
<div class="example">
<pre class="example">gprof <var>options</var> [<var>executable-file</var> [<var>profile-data-files</var>&hellip;]] [&gt; <var>outfile</var>]
</pre></div>

<p>Here square-brackets indicate optional arguments.
</p>
<p>If you omit the executable file name, the file <samp>a.out</samp> is used.  If
you give no profile data file name, the file <samp>gmon.out</samp> is used.  If
any file is not in the proper format, or if the profile data file does not
appear to belong to the executable file, an error message is printed.
</p>
<p>You can give more than one profile data file by entering all their names
after the executable file name; then the statistics in all the data files
are summed together.
</p>
<p>The order of these options does not matter.
</p>
<table class="menu" border="0" cellspacing="0">
<tr><td align="left" valign="top">&bull; <a href="#Output-Options" accesskey="1">Output Options</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Controlling <code>gprof</code>&rsquo;s output style
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Analysis-Options" accesskey="2">Analysis Options</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Controlling how <code>gprof</code> analyzes its data
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Miscellaneous-Options" accesskey="3">Miscellaneous Options</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Deprecated-Options" accesskey="4">Deprecated Options</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Options you no longer need to use, but which
                            have been retained for compatibility
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Symspecs" accesskey="5">Symspecs</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Specifying functions to include or exclude
</td></tr>
</table>

<hr>
<span id="Output-Options"></span><div class="header">
<p>
Next: <a href="#Analysis-Options" accesskey="n" rel="next">Analysis Options</a>, Up: <a href="#Invoking" accesskey="u" rel="up">Invoking</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Output-Options-1"></span><h3 class="section">4.1 Output Options</h3>

<p>These options specify which of several output formats
<code>gprof</code> should produce.
</p>
<p>Many of these options take an optional <em>symspec</em> to specify
functions to be included or excluded.  These options can be
specified multiple times, with different symspecs, to include
or exclude sets of symbols.  See <a href="#Symspecs">Symspecs</a>.
</p>
<p>Specifying any of these options overrides the default (&lsquo;<samp>-p -q</samp>&rsquo;),
which prints a flat profile and call graph analysis
for all functions.
</p>
<dl compact="compact">
<dt><code>-A[<var>symspec</var>]</code></dt>
<dt><code>--annotated-source[=<var>symspec</var>]</code></dt>
<dd><p>The &lsquo;<samp>-A</samp>&rsquo; option causes <code>gprof</code> to print annotated source code.
If <var>symspec</var> is specified, print output only for matching symbols.
See <a href="#Annotated-Source">The Annotated Source Listing</a>.
</p>
</dd>
<dt><code>-b</code></dt>
<dt><code>--brief</code></dt>
<dd><p>If the &lsquo;<samp>-b</samp>&rsquo; option is given, <code>gprof</code> doesn&rsquo;t print the
verbose blurbs that try to explain the meaning of all of the fields in
the tables.  This is useful if you intend to print out the output, or
are tired of seeing the blurbs.
</p>
</dd>
<dt><code>-B</code></dt>
<dd><p>The &lsquo;<samp>-B</samp>&rsquo; option causes <code>gprof</code> to print the call graph analysis.
</p>
</dd>
<dt><code>-C[<var>symspec</var>]</code></dt>
<dt><code>--exec-counts[=<var>symspec</var>]</code></dt>
<dd><p>The &lsquo;<samp>-C</samp>&rsquo; option causes <code>gprof</code> to
print a tally of functions and the number of times each was called.
If <var>symspec</var> is specified, print tally only for matching symbols.
</p>
<p>If the profile data file contains basic-block count records, specifying
the &lsquo;<samp>-l</samp>&rsquo; option, along with &lsquo;<samp>-C</samp>&rsquo;, will cause basic-block
execution counts to be tallied and displayed.
</p>
</dd>
<dt><code>-i</code></dt>
<dt><code>--file-info</code></dt>
<dd><p>The &lsquo;<samp>-i</samp>&rsquo; option causes <code>gprof</code> to display summary information
about the profile data file(s) and then exit.  The number of histogram,
call graph, and basic-block count records is displayed.
</p>
</dd>
<dt><code>-I <var>dirs</var></code></dt>
<dt><code>--directory-path=<var>dirs</var></code></dt>
<dd><p>The &lsquo;<samp>-I</samp>&rsquo; option specifies a list of search directories in
which to find source files.  Environment variable <var>GPROF_PATH</var>
can also be used to convey this information.
Used mostly for annotated source output.
</p>
</dd>
<dt><code>-J[<var>symspec</var>]</code></dt>
<dt><code>--no-annotated-source[=<var>symspec</var>]</code></dt>
<dd><p>The &lsquo;<samp>-J</samp>&rsquo; option causes <code>gprof</code> not to
print annotated source code.
If <var>symspec</var> is specified, <code>gprof</code> prints annotated source,
but excludes matching symbols.
</p>
</dd>
<dt><code>-L</code></dt>
<dt><code>--print-path</code></dt>
<dd><p>Normally, source filenames are printed with the path
component suppressed.  The &lsquo;<samp>-L</samp>&rsquo; option causes <code>gprof</code>
to print the full pathname of
source filenames, which is determined
from symbolic debugging information in the image file
and is relative to the directory in which the compiler
was invoked.
</p>
</dd>
<dt><code>-p[<var>symspec</var>]</code></dt>
<dt><code>--flat-profile[=<var>symspec</var>]</code></dt>
<dd><p>The &lsquo;<samp>-p</samp>&rsquo; option causes <code>gprof</code> to print a flat profile.
If <var>symspec</var> is specified, print flat profile only for matching symbols.
See <a href="#Flat-Profile">The Flat Profile</a>.
</p>
</dd>
<dt><code>-P[<var>symspec</var>]</code></dt>
<dt><code>--no-flat-profile[=<var>symspec</var>]</code></dt>
<dd><p>The &lsquo;<samp>-P</samp>&rsquo; option causes <code>gprof</code> to suppress printing a flat profile.
If <var>symspec</var> is specified, <code>gprof</code> prints a flat profile,
but excludes matching symbols.
</p>
</dd>
<dt><code>-q[<var>symspec</var>]</code></dt>
<dt><code>--graph[=<var>symspec</var>]</code></dt>
<dd><p>The &lsquo;<samp>-q</samp>&rsquo; option causes <code>gprof</code> to print the call graph analysis.
If <var>symspec</var> is specified, print call graph only for matching symbols
and their children.
See <a href="#Call-Graph">The Call Graph</a>.
</p>
</dd>
<dt><code>-Q[<var>symspec</var>]</code></dt>
<dt><code>--no-graph[=<var>symspec</var>]</code></dt>
<dd><p>The &lsquo;<samp>-Q</samp>&rsquo; option causes <code>gprof</code> to suppress printing the
call graph.
If <var>symspec</var> is specified, <code>gprof</code> prints a call graph,
but excludes matching symbols.
</p>
</dd>
<dt><code>-t</code></dt>
<dt><code>--table-length=<var>num</var></code></dt>
<dd><p>The &lsquo;<samp>-t</samp>&rsquo; option causes the <var>num</var> most active source lines in
each source file to be listed when source annotation is enabled.  The
default is 10.
</p>
</dd>
<dt><code>-y</code></dt>
<dt><code>--separate-files</code></dt>
<dd><p>This option affects annotated source output only.
Normally, <code>gprof</code> prints annotated source files
to standard-output.  If this option is specified,
annotated source for a file named <samp>path/<var>filename</var></samp>
is generated in the file <samp><var>filename</var>-ann</samp>.  If the underlying
file system would truncate <samp><var>filename</var>-ann</samp> so that it
overwrites the original <samp><var>filename</var></samp>, <code>gprof</code> generates
annotated source in the file <samp><var>filename</var>.ann</samp> instead (if the
original file name has an extension, that extension is <em>replaced</em>
with <samp>.ann</samp>).
</p>
</dd>
<dt><code>-Z[<var>symspec</var>]</code></dt>
<dt><code>--no-exec-counts[=<var>symspec</var>]</code></dt>
<dd><p>The &lsquo;<samp>-Z</samp>&rsquo; option causes <code>gprof</code> not to
print a tally of functions and the number of times each was called.
If <var>symspec</var> is specified, print tally, but exclude matching symbols.
</p>
</dd>
<dt><code>-r</code></dt>
<dt><code>--function-ordering</code></dt>
<dd><p>The &lsquo;<samp>--function-ordering</samp>&rsquo; option causes <code>gprof</code> to print a
suggested function ordering for the program based on profiling data.
This option suggests an ordering which may improve paging, tlb and
cache behavior for the program on systems which support arbitrary
ordering of functions in an executable.
</p>
<p>The exact details of how to force the linker to place functions
in a particular order is system dependent and out of the scope of this
manual.
</p>
</dd>
<dt><code>-R <var>map_file</var></code></dt>
<dt><code>--file-ordering <var>map_file</var></code></dt>
<dd><p>The &lsquo;<samp>--file-ordering</samp>&rsquo; option causes <code>gprof</code> to print a
suggested .o link line ordering for the program based on profiling data.
This option suggests an ordering which may improve paging, tlb and
cache behavior for the program on systems which do not support arbitrary
ordering of functions in an executable.
</p>
<p>Use of the &lsquo;<samp>-a</samp>&rsquo; argument is highly recommended with this option.
</p>
<p>The <var>map_file</var> argument is a pathname to a file which provides
function name to object file mappings.  The format of the file is similar to
the output of the program <code>nm</code>.
</p>
<div class="example">
<pre class="example">c-parse.o:00000000 T yyparse
c-parse.o:00000004 C yyerrflag
c-lang.o:00000000 T maybe_objc_method_name
c-lang.o:00000000 T print_lang_statistics
c-lang.o:00000000 T recognize_objc_keyword
c-decl.o:00000000 T print_lang_identifier
c-decl.o:00000000 T print_lang_type
&hellip;

</pre></div>

<p>To create a <var>map_file</var> with <small>GNU</small> <code>nm</code>, type a command like
<kbd>nm --extern-only --defined-only -v --print-file-name program-name</kbd>.
</p>
</dd>
<dt><code>-T</code></dt>
<dt><code>--traditional</code></dt>
<dd><p>The &lsquo;<samp>-T</samp>&rsquo; option causes <code>gprof</code> to print its output in
&ldquo;traditional&rdquo; BSD style.
</p>
</dd>
<dt><code>-w <var>width</var></code></dt>
<dt><code>--width=<var>width</var></code></dt>
<dd><p>Sets width of output lines to <var>width</var>.
Currently only used when printing the function index at the bottom
of the call graph.
</p>
</dd>
<dt><code>-x</code></dt>
<dt><code>--all-lines</code></dt>
<dd><p>This option affects annotated source output only.
By default, only the lines at the beginning of a basic-block
are annotated.  If this option is specified, every line in
a basic-block is annotated by repeating the annotation for the
first line.  This behavior is similar to <code>tcov</code>&rsquo;s &lsquo;<samp>-a</samp>&rsquo;.
</p>
</dd>
<dt><code>--demangle[=<var>style</var>]</code></dt>
<dt><code>--no-demangle</code></dt>
<dd><p>These options control whether C++ symbol names should be demangled when
printing output.  The default is to demangle symbols.  The
<code>--no-demangle</code> option may be used to turn off demangling. Different
compilers have different mangling styles.  The optional demangling style
argument can be used to choose an appropriate demangling style for your
compiler.
</p></dd>
</dl>

<hr>
<span id="Analysis-Options"></span><div class="header">
<p>
Next: <a href="#Miscellaneous-Options" accesskey="n" rel="next">Miscellaneous Options</a>, Previous: <a href="#Output-Options" accesskey="p" rel="prev">Output Options</a>, Up: <a href="#Invoking" accesskey="u" rel="up">Invoking</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Analysis-Options-1"></span><h3 class="section">4.2 Analysis Options</h3>

<dl compact="compact">
<dt><code>-a</code></dt>
<dt><code>--no-static</code></dt>
<dd><p>The &lsquo;<samp>-a</samp>&rsquo; option causes <code>gprof</code> to suppress the printing of
statically declared (private) functions.  (These are functions whose
names are not listed as global, and which are not visible outside the
file/function/block where they were defined.)  Time spent in these
functions, calls to/from them, etc., will all be attributed to the
function that was loaded directly before it in the executable file.
This option affects both the flat profile and the call graph.
</p>
</dd>
<dt><code>-c</code></dt>
<dt><code>--static-call-graph</code></dt>
<dd><p>The &lsquo;<samp>-c</samp>&rsquo; option causes the call graph of the program to be
augmented by a heuristic which examines the text space of the object
file and identifies function calls in the binary machine code.
Since normal call graph records are only generated when functions are
entered, this option identifies children that could have been called,
but never were.  Calls to functions that were not compiled with
profiling enabled are also identified, but only if symbol table
entries are present for them.
Calls to dynamic library routines are typically <em>not</em> found
by this option.
Parents or children identified via this heuristic
are indicated in the call graph with call counts of &lsquo;<samp>0</samp>&rsquo;.
</p>
</dd>
<dt><code>-D</code></dt>
<dt><code>--ignore-non-functions</code></dt>
<dd><p>The &lsquo;<samp>-D</samp>&rsquo; option causes <code>gprof</code> to ignore symbols which
are not known to be functions.  This option will give more accurate
profile data on systems where it is supported (Solaris and HPUX for
example).
</p>
</dd>
<dt><code>-k <var>from</var>/<var>to</var></code></dt>
<dd><p>The &lsquo;<samp>-k</samp>&rsquo; option allows you to delete from the call graph any arcs from
symbols matching symspec <var>from</var> to those matching symspec <var>to</var>.
</p>
</dd>
<dt><code>-l</code></dt>
<dt><code>--line</code></dt>
<dd><p>The &lsquo;<samp>-l</samp>&rsquo; option enables line-by-line profiling, which causes
histogram hits to be charged to individual source code lines,
instead of functions.  This feature only works with programs compiled
by older versions of the <code>gcc</code> compiler.  Newer versions of
<code>gcc</code> are designed to work with the <code>gcov</code> tool instead.
</p>
<p>If the program was compiled with basic-block counting enabled,
this option will also identify how many times each line of
code was executed.
While line-by-line profiling can help isolate where in a large function
a program is spending its time, it also significantly increases
the running time of <code>gprof</code>, and magnifies statistical
inaccuracies.
See <a href="#Sampling-Error">Statistical Sampling Error</a>.
</p>
</dd>
<dt><code>--inline-file-names</code></dt>
<dd><p>This option causes <code>gprof</code> to print the source file after each
symbol in both the flat profile and the call graph. The full path to the
file is printed if used with the &lsquo;<samp>-L</samp>&rsquo; option.
</p>
</dd>
<dt><code>-m <var>num</var></code></dt>
<dt><code>--min-count=<var>num</var></code></dt>
<dd><p>This option affects execution count output only.
Symbols that are executed less than <var>num</var> times are suppressed.
</p>
</dd>
<dt><code>-n<var>symspec</var></code></dt>
<dt><code>--time=<var>symspec</var></code></dt>
<dd><p>The &lsquo;<samp>-n</samp>&rsquo; option causes <code>gprof</code>, in its call graph analysis,
to only propagate times for symbols matching <var>symspec</var>.
</p>
</dd>
<dt><code>-N<var>symspec</var></code></dt>
<dt><code>--no-time=<var>symspec</var></code></dt>
<dd><p>The &lsquo;<samp>-n</samp>&rsquo; option causes <code>gprof</code>, in its call graph analysis,
not to propagate times for symbols matching <var>symspec</var>.
</p>
</dd>
<dt><code>-S<var>filename</var></code></dt>
<dt><code>--external-symbol-table=<var>filename</var></code></dt>
<dd><p>The &lsquo;<samp>-S</samp>&rsquo; option causes <code>gprof</code> to read an external symbol table
file, such as <samp>/proc/kallsyms</samp>, rather than read the symbol table
from the given object file (the default is <code>a.out</code>). This is useful
for profiling kernel modules.
</p>
</dd>
<dt><code>-z</code></dt>
<dt><code>--display-unused-functions</code></dt>
<dd><p>If you give the &lsquo;<samp>-z</samp>&rsquo; option, <code>gprof</code> will mention all
functions in the flat profile, even those that were never called, and
that had no time spent in them.  This is useful in conjunction with the
&lsquo;<samp>-c</samp>&rsquo; option for discovering which routines were never called.
</p>
</dd>
</dl>

<hr>
<span id="Miscellaneous-Options"></span><div class="header">
<p>
Next: <a href="#Deprecated-Options" accesskey="n" rel="next">Deprecated Options</a>, Previous: <a href="#Analysis-Options" accesskey="p" rel="prev">Analysis Options</a>, Up: <a href="#Invoking" accesskey="u" rel="up">Invoking</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Miscellaneous-Options-1"></span><h3 class="section">4.3 Miscellaneous Options</h3>

<dl compact="compact">
<dt><code>-d[<var>num</var>]</code></dt>
<dt><code>--debug[=<var>num</var>]</code></dt>
<dd><p>The &lsquo;<samp>-d <var>num</var></samp>&rsquo; option specifies debugging options.
If <var>num</var> is not specified, enable all debugging.
See <a href="#Debugging">Debugging <code>gprof</code></a>.
</p>
</dd>
<dt><code>-h</code></dt>
<dt><code>--help</code></dt>
<dd><p>The &lsquo;<samp>-h</samp>&rsquo; option prints command line usage.
</p>
</dd>
<dt><code>-O<var>name</var></code></dt>
<dt><code>--file-format=<var>name</var></code></dt>
<dd><p>Selects the format of the profile data files.  Recognized formats are
&lsquo;<samp>auto</samp>&rsquo; (the default), &lsquo;<samp>bsd</samp>&rsquo;, &lsquo;<samp>4.4bsd</samp>&rsquo;, &lsquo;<samp>magic</samp>&rsquo;, and
&lsquo;<samp>prof</samp>&rsquo; (not yet supported).
</p>
</dd>
<dt><code>-s</code></dt>
<dt><code>--sum</code></dt>
<dd><p>The &lsquo;<samp>-s</samp>&rsquo; option causes <code>gprof</code> to summarize the information
in the profile data files it read in, and write out a profile data
file called <samp>gmon.sum</samp>, which contains all the information from
the profile data files that <code>gprof</code> read in.  The file <samp>gmon.sum</samp>
may be one of the specified input files; the effect of this is to
merge the data in the other input files into <samp>gmon.sum</samp>.
</p>
<p>Eventually you can run <code>gprof</code> again without &lsquo;<samp>-s</samp>&rsquo; to analyze the
cumulative data in the file <samp>gmon.sum</samp>.
</p>
</dd>
<dt><code>-v</code></dt>
<dt><code>--version</code></dt>
<dd><p>The &lsquo;<samp>-v</samp>&rsquo; flag causes <code>gprof</code> to print the current version
number, and then exit.
</p>
</dd>
</dl>

<hr>
<span id="Deprecated-Options"></span><div class="header">
<p>
Next: <a href="#Symspecs" accesskey="n" rel="next">Symspecs</a>, Previous: <a href="#Miscellaneous-Options" accesskey="p" rel="prev">Miscellaneous Options</a>, Up: <a href="#Invoking" accesskey="u" rel="up">Invoking</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Deprecated-Options-1"></span><h3 class="section">4.4 Deprecated Options</h3>

<p>These options have been replaced with newer versions that use symspecs.
</p>
<dl compact="compact">
<dt><code>-e <var>function_name</var></code></dt>
<dd><p>The &lsquo;<samp>-e <var>function</var></samp>&rsquo; option tells <code>gprof</code> to not print
information about the function <var>function_name</var> (and its
children&hellip;) in the call graph.  The function will still be listed
as a child of any functions that call it, but its index number will be
shown as &lsquo;<samp>[not printed]</samp>&rsquo;.  More than one &lsquo;<samp>-e</samp>&rsquo; option may be
given; only one <var>function_name</var> may be indicated with each &lsquo;<samp>-e</samp>&rsquo;
option.
</p>
</dd>
<dt><code>-E <var>function_name</var></code></dt>
<dd><p>The <code>-E <var>function</var></code> option works like the <code>-e</code> option, but
time spent in the function (and children who were not called from
anywhere else), will not be used to compute the percentages-of-time for
the call graph.  More than one &lsquo;<samp>-E</samp>&rsquo; option may be given; only one
<var>function_name</var> may be indicated with each &lsquo;<samp>-E</samp>&rsquo; option.
</p>
</dd>
<dt><code>-f <var>function_name</var></code></dt>
<dd><p>The &lsquo;<samp>-f <var>function</var></samp>&rsquo; option causes <code>gprof</code> to limit the
call graph to the function <var>function_name</var> and its children (and
their children&hellip;).  More than one &lsquo;<samp>-f</samp>&rsquo; option may be given;
only one <var>function_name</var> may be indicated with each &lsquo;<samp>-f</samp>&rsquo;
option.
</p>
</dd>
<dt><code>-F <var>function_name</var></code></dt>
<dd><p>The &lsquo;<samp>-F <var>function</var></samp>&rsquo; option works like the <code>-f</code> option, but
only time spent in the function and its children (and their
children&hellip;) will be used to determine total-time and
percentages-of-time for the call graph.  More than one &lsquo;<samp>-F</samp>&rsquo; option
may be given; only one <var>function_name</var> may be indicated with each
&lsquo;<samp>-F</samp>&rsquo; option.  The &lsquo;<samp>-F</samp>&rsquo; option overrides the &lsquo;<samp>-E</samp>&rsquo; option.
</p>
</dd>
</dl>


<p>Note that only one function can be specified with each <code>-e</code>,
<code>-E</code>, <code>-f</code> or <code>-F</code> option.  To specify more than one
function, use multiple options.  For example, this command:
</p>
<div class="example">
<pre class="example">gprof -e boring -f foo -f bar myprogram &gt; gprof.output
</pre></div>

<p>lists in the call graph all functions that were reached from either
<code>foo</code> or <code>bar</code> and were not reachable from <code>boring</code>.
</p>
<hr>
<span id="Symspecs"></span><div class="header">
<p>
Previous: <a href="#Deprecated-Options" accesskey="p" rel="prev">Deprecated Options</a>, Up: <a href="#Invoking" accesskey="u" rel="up">Invoking</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Symspecs-1"></span><h3 class="section">4.5 Symspecs</h3>

<p>Many of the output options allow functions to be included or excluded
using <em>symspecs</em> (symbol specifications), which observe the
following syntax:
</p>
<div class="example">
<pre class="example">  filename_containing_a_dot
| funcname_not_containing_a_dot
| linenumber
| ( [ any_filename ] `:' ( any_funcname | linenumber ) )
</pre></div>

<p>Here are some sample symspecs:
</p>
<dl compact="compact">
<dt>&lsquo;<samp>main.c</samp>&rsquo;</dt>
<dd><p>Selects everything in file <samp>main.c</samp>&mdash;the
dot in the string tells <code>gprof</code> to interpret
the string as a filename, rather than as
a function name.  To select a file whose
name does not contain a dot, a trailing colon
should be specified.  For example, &lsquo;<samp>odd:</samp>&rsquo; is
interpreted as the file named <samp>odd</samp>.
</p>
</dd>
<dt>&lsquo;<samp>main</samp>&rsquo;</dt>
<dd><p>Selects all functions named &lsquo;<samp>main</samp>&rsquo;.
</p>
<p>Note that there may be multiple instances of the same function name
because some of the definitions may be local (i.e., static).  Unless a
function name is unique in a program, you must use the colon notation
explained below to specify a function from a specific source file.
</p>
<p>Sometimes, function names contain dots.  In such cases, it is necessary
to add a leading colon to the name.  For example, &lsquo;<samp>:.mul</samp>&rsquo; selects
function &lsquo;<samp>.mul</samp>&rsquo;.
</p>
<p>In some object file formats, symbols have a leading underscore.
<code>gprof</code> will normally not print these underscores.  When you name a
symbol in a symspec, you should type it exactly as <code>gprof</code> prints
it in its output.  For example, if the compiler produces a symbol
&lsquo;<samp>_main</samp>&rsquo; from your <code>main</code> function, <code>gprof</code> still prints
it as &lsquo;<samp>main</samp>&rsquo; in its output, so you should use &lsquo;<samp>main</samp>&rsquo; in
symspecs.
</p>
</dd>
<dt>&lsquo;<samp>main.c:main</samp>&rsquo;</dt>
<dd><p>Selects function &lsquo;<samp>main</samp>&rsquo; in file <samp>main.c</samp>.
</p>
</dd>
<dt>&lsquo;<samp>main.c:134</samp>&rsquo;</dt>
<dd><p>Selects line 134 in file <samp>main.c</samp>.
</p></dd>
</dl>

<hr>
<span id="Output"></span><div class="header">
<p>
Next: <a href="#Inaccuracy" accesskey="n" rel="next">Inaccuracy</a>, Previous: <a href="#Invoking" accesskey="p" rel="prev">Invoking</a>, Up: <a href="#Top" accesskey="u" rel="up">Top</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Interpreting-gprof_0027s-Output"></span><h2 class="chapter">5 Interpreting <code>gprof</code>&rsquo;s Output</h2>

<p><code>gprof</code> can produce several different output styles, the
most important of which are described below.  The simplest output
styles (file information, execution count, and function and file ordering)
are not described here, but are documented with the respective options
that trigger them.
See <a href="#Output-Options">Output Options</a>.
</p>
<table class="menu" border="0" cellspacing="0">
<tr><td align="left" valign="top">&bull; <a href="#Flat-Profile" accesskey="1">Flat Profile</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">The flat profile shows how much time was spent
                            executing directly in each function.
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Call-Graph" accesskey="2">Call Graph</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">The call graph shows which functions called which
                            others, and how much time each function used
                            when its subroutine calls are included.
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Line_002dby_002dline" accesskey="3">Line-by-line</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top"><code>gprof</code> can analyze individual source code lines
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Annotated-Source" accesskey="4">Annotated Source</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">The annotated source listing displays source code
                            labeled with execution counts
</td></tr>
</table>


<hr>
<span id="Flat-Profile"></span><div class="header">
<p>
Next: <a href="#Call-Graph" accesskey="n" rel="next">Call Graph</a>, Up: <a href="#Output" accesskey="u" rel="up">Output</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="The-Flat-Profile"></span><h3 class="section">5.1 The Flat Profile</h3>
<span id="index-flat-profile"></span>

<p>The <em>flat profile</em> shows the total amount of time your program
spent executing each function.  Unless the &lsquo;<samp>-z</samp>&rsquo; option is given,
functions with no apparent time spent in them, and no apparent calls
to them, are not mentioned.  Note that if a function was not compiled
for profiling, and didn&rsquo;t run long enough to show up on the program
counter histogram, it will be indistinguishable from a function that
was never called.
</p>
<p>This is part of a flat profile for a small program:
</p>
<div class="example">
<pre class="example">Flat profile:

Each sample counts as 0.01 seconds.
  %   cumulative   self              self     total
 time   seconds   seconds    calls  ms/call  ms/call  name
 33.34      0.02     0.02     7208     0.00     0.00  open
 16.67      0.03     0.01      244     0.04     0.12  offtime
 16.67      0.04     0.01        8     1.25     1.25  memccpy
 16.67      0.05     0.01        7     1.43     1.43  write
 16.67      0.06     0.01                             mcount
  0.00      0.06     0.00      236     0.00     0.00  tzset
  0.00      0.06     0.00      192     0.00     0.00  tolower
  0.00      0.06     0.00       47     0.00     0.00  strlen
  0.00      0.06     0.00       45     0.00     0.00  strchr
  0.00      0.06     0.00        1     0.00    50.00  main
  0.00      0.06     0.00        1     0.00     0.00  memcpy
  0.00      0.06     0.00        1     0.00    10.11  print
  0.00      0.06     0.00        1     0.00     0.00  profil
  0.00      0.06     0.00        1     0.00    50.00  report
&hellip;
</pre></div>

<p>The functions are sorted first by decreasing run-time spent in them,
then by decreasing number of calls, then alphabetically by name.  The
functions &lsquo;<samp>mcount</samp>&rsquo; and &lsquo;<samp>profil</samp>&rsquo; are part of the profiling
apparatus and appear in every flat profile; their time gives a measure of
the amount of overhead due to profiling.
</p>
<p>Just before the column headers, a statement appears indicating
how much time each sample counted as.
This <em>sampling period</em> estimates the margin of error in each of the time
figures.  A time figure that is not much larger than this is not
reliable.  In this example, each sample counted as 0.01 seconds,
suggesting a 100 Hz sampling rate.
The program&rsquo;s total execution time was 0.06
seconds, as indicated by the &lsquo;<samp>cumulative seconds</samp>&rsquo; field.  Since
each sample counted for 0.01 seconds, this means only six samples
were taken during the run.  Two of the samples occurred while the
program was in the &lsquo;<samp>open</samp>&rsquo; function, as indicated by the
&lsquo;<samp>self seconds</samp>&rsquo; field.  Each of the other four samples
occurred one each in &lsquo;<samp>offtime</samp>&rsquo;, &lsquo;<samp>memccpy</samp>&rsquo;, &lsquo;<samp>write</samp>&rsquo;,
and &lsquo;<samp>mcount</samp>&rsquo;.
Since only six samples were taken, none of these values can
be regarded as particularly reliable.
In another run,
the &lsquo;<samp>self seconds</samp>&rsquo; field for
&lsquo;<samp>mcount</samp>&rsquo; might well be &lsquo;<samp>0.00</samp>&rsquo; or &lsquo;<samp>0.02</samp>&rsquo;.
See <a href="#Sampling-Error">Statistical Sampling Error</a>,
for a complete discussion.
</p>
<p>The remaining functions in the listing (those whose
&lsquo;<samp>self seconds</samp>&rsquo; field is &lsquo;<samp>0.00</samp>&rsquo;) didn&rsquo;t appear
in the histogram samples at all.  However, the call graph
indicated that they were called, so therefore they are listed,
sorted in decreasing order by the &lsquo;<samp>calls</samp>&rsquo; field.
Clearly some time was spent executing these functions,
but the paucity of histogram samples prevents any
determination of how much time each took.
</p>
<p>Here is what the fields in each line mean:
</p>
<dl compact="compact">
<dt><code>% time</code></dt>
<dd><p>This is the percentage of the total execution time your program spent
in this function.  These should all add up to 100%.
</p>
</dd>
<dt><code>cumulative seconds</code></dt>
<dd><p>This is the cumulative total number of seconds the computer spent
executing this functions, plus the time spent in all the functions
above this one in this table.
</p>
</dd>
<dt><code>self seconds</code></dt>
<dd><p>This is the number of seconds accounted for by this function alone.
The flat profile listing is sorted first by this number.
</p>
</dd>
<dt><code>calls</code></dt>
<dd><p>This is the total number of times the function was called.  If the
function was never called, or the number of times it was called cannot
be determined (probably because the function was not compiled with
profiling enabled), the <em>calls</em> field is blank.
</p>
</dd>
<dt><code>self ms/call</code></dt>
<dd><p>This represents the average number of milliseconds spent in this
function per call, if this function is profiled.  Otherwise, this field
is blank for this function.
</p>
</dd>
<dt><code>total ms/call</code></dt>
<dd><p>This represents the average number of milliseconds spent in this
function and its descendants per call, if this function is profiled.
Otherwise, this field is blank for this function.
This is the only field in the flat profile that uses call graph analysis.
</p>
</dd>
<dt><code>name</code></dt>
<dd><p>This is the name of the function.   The flat profile is sorted by this
field alphabetically after the <em>self seconds</em> and <em>calls</em>
fields are sorted.
</p></dd>
</dl>

<hr>
<span id="Call-Graph"></span><div class="header">
<p>
Next: <a href="#Line_002dby_002dline" accesskey="n" rel="next">Line-by-line</a>, Previous: <a href="#Flat-Profile" accesskey="p" rel="prev">Flat Profile</a>, Up: <a href="#Output" accesskey="u" rel="up">Output</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="The-Call-Graph"></span><h3 class="section">5.2 The Call Graph</h3>
<span id="index-call-graph"></span>

<p>The <em>call graph</em> shows how much time was spent in each function
and its children.  From this information, you can find functions that,
while they themselves may not have used much time, called other
functions that did use unusual amounts of time.
</p>
<p>Here is a sample call from a small program.  This call came from the
same <code>gprof</code> run as the flat profile example in the previous
section.
</p>
<div class="example">
<pre class="example">granularity: each sample hit covers 2 byte(s) for 20.00% of 0.05 seconds

index % time    self  children    called     name
                                                 &lt;spontaneous&gt;
[1]    100.0    0.00    0.05                 start [1]
                0.00    0.05       1/1           main [2]
                0.00    0.00       1/2           on_exit [28]
                0.00    0.00       1/1           exit [59]
-----------------------------------------------
                0.00    0.05       1/1           start [1]
[2]    100.0    0.00    0.05       1         main [2]
                0.00    0.05       1/1           report [3]
-----------------------------------------------
                0.00    0.05       1/1           main [2]
[3]    100.0    0.00    0.05       1         report [3]
                0.00    0.03       8/8           timelocal [6]
                0.00    0.01       1/1           print [9]
                0.00    0.01       9/9           fgets [12]
                0.00    0.00      12/34          strncmp &lt;cycle 1&gt; [40]
                0.00    0.00       8/8           lookup [20]
                0.00    0.00       1/1           fopen [21]
                0.00    0.00       8/8           chewtime [24]
                0.00    0.00       8/16          skipspace [44]
-----------------------------------------------
[4]     59.8    0.01        0.02       8+472     &lt;cycle 2 as a whole&gt; [4]
                0.01        0.02     244+260         offtime &lt;cycle 2&gt; [7]
                0.00        0.00     236+1           tzset &lt;cycle 2&gt; [26]
-----------------------------------------------
</pre></div>

<p>The lines full of dashes divide this table into <em>entries</em>, one for each
function.  Each entry has one or more lines.
</p>
<p>In each entry, the primary line is the one that starts with an index number
in square brackets.  The end of this line says which function the entry is
for.  The preceding lines in the entry describe the callers of this
function and the following lines describe its subroutines (also called
<em>children</em> when we speak of the call graph).
</p>
<p>The entries are sorted by time spent in the function and its subroutines.
</p>
<p>The internal profiling function <code>mcount</code> (see <a href="#Flat-Profile">The
Flat Profile</a>) is never mentioned in the call graph.
</p>
<table class="menu" border="0" cellspacing="0">
<tr><td align="left" valign="top">&bull; <a href="#Primary" accesskey="1">Primary</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Details of the primary line&rsquo;s contents.
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Callers" accesskey="2">Callers</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Details of caller-lines&rsquo; contents.
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Subroutines" accesskey="3">Subroutines</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Details of subroutine-lines&rsquo; contents.
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Cycles" accesskey="4">Cycles</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">When there are cycles of recursion,
                   such as <code>a</code> calls <code>b</code> calls <code>a</code>&hellip;
</td></tr>
</table>

<hr>
<span id="Primary"></span><div class="header">
<p>
Next: <a href="#Callers" accesskey="n" rel="next">Callers</a>, Up: <a href="#Call-Graph" accesskey="u" rel="up">Call Graph</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="The-Primary-Line"></span><h4 class="subsection">5.2.1 The Primary Line</h4>

<p>The <em>primary line</em> in a call graph entry is the line that
describes the function which the entry is about and gives the overall
statistics for this function.
</p>
<p>For reference, we repeat the primary line from the entry for function
<code>report</code> in our main example, together with the heading line that
shows the names of the fields:
</p>
<div class="example">
<pre class="example">index  % time    self  children called     name
&hellip;
[3]    100.0    0.00    0.05       1         report [3]
</pre></div>

<p>Here is what the fields in the primary line mean:
</p>
<dl compact="compact">
<dt><code>index</code></dt>
<dd><p>Entries are numbered with consecutive integers.  Each function
therefore has an index number, which appears at the beginning of its
primary line.
</p>
<p>Each cross-reference to a function, as a caller or subroutine of
another, gives its index number as well as its name.  The index number
guides you if you wish to look for the entry for that function.
</p>
</dd>
<dt><code>% time</code></dt>
<dd><p>This is the percentage of the total time that was spent in this
function, including time spent in subroutines called from this
function.
</p>
<p>The time spent in this function is counted again for the callers of
this function.  Therefore, adding up these percentages is meaningless.
</p>
</dd>
<dt><code>self</code></dt>
<dd><p>This is the total amount of time spent in this function.  This
should be identical to the number printed in the <code>seconds</code> field
for this function in the flat profile.
</p>
</dd>
<dt><code>children</code></dt>
<dd><p>This is the total amount of time spent in the subroutine calls made by
this function.  This should be equal to the sum of all the <code>self</code>
and <code>children</code> entries of the children listed directly below this
function.
</p>
</dd>
<dt><code>called</code></dt>
<dd><p>This is the number of times the function was called.
</p>
<p>If the function called itself recursively, there are two numbers,
separated by a &lsquo;<samp>+</samp>&rsquo;.  The first number counts non-recursive calls,
and the second counts recursive calls.
</p>
<p>In the example above, the function <code>report</code> was called once from
<code>main</code>.
</p>
</dd>
<dt><code>name</code></dt>
<dd><p>This is the name of the current function.  The index number is
repeated after it.
</p>
<p>If the function is part of a cycle of recursion, the cycle number is
printed between the function&rsquo;s name and the index number
(see <a href="#Cycles">How Mutually Recursive Functions Are Described</a>).
For example, if function <code>gnurr</code> is part of
cycle number one, and has index number twelve, its primary line would
be end like this:
</p>
<div class="example">
<pre class="example">gnurr &lt;cycle 1&gt; [12]
</pre></div>
</dd>
</dl>

<hr>
<span id="Callers"></span><div class="header">
<p>
Next: <a href="#Subroutines" accesskey="n" rel="next">Subroutines</a>, Previous: <a href="#Primary" accesskey="p" rel="prev">Primary</a>, Up: <a href="#Call-Graph" accesskey="u" rel="up">Call Graph</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Lines-for-a-Function_0027s-Callers"></span><h4 class="subsection">5.2.2 Lines for a Function&rsquo;s Callers</h4>

<p>A function&rsquo;s entry has a line for each function it was called by.
These lines&rsquo; fields correspond to the fields of the primary line, but
their meanings are different because of the difference in context.
</p>
<p>For reference, we repeat two lines from the entry for the function
<code>report</code>, the primary line and one caller-line preceding it, together
with the heading line that shows the names of the fields:
</p>
<div class="example">
<pre class="example">index  % time    self  children called     name
&hellip;
                0.00    0.05       1/1           main [2]
[3]    100.0    0.00    0.05       1         report [3]
</pre></div>

<p>Here are the meanings of the fields in the caller-line for <code>report</code>
called from <code>main</code>:
</p>
<dl compact="compact">
<dt><code>self</code></dt>
<dd><p>An estimate of the amount of time spent in <code>report</code> itself when it was
called from <code>main</code>.
</p>
</dd>
<dt><code>children</code></dt>
<dd><p>An estimate of the amount of time spent in subroutines of <code>report</code>
when <code>report</code> was called from <code>main</code>.
</p>
<p>The sum of the <code>self</code> and <code>children</code> fields is an estimate
of the amount of time spent within calls to <code>report</code> from <code>main</code>.
</p>
</dd>
<dt><code>called</code></dt>
<dd><p>Two numbers: the number of times <code>report</code> was called from <code>main</code>,
followed by the total number of non-recursive calls to <code>report</code> from
all its callers.
</p>
</dd>
<dt><code>name and index number</code></dt>
<dd><p>The name of the caller of <code>report</code> to which this line applies,
followed by the caller&rsquo;s index number.
</p>
<p>Not all functions have entries in the call graph; some
options to <code>gprof</code> request the omission of certain functions.
When a caller has no entry of its own, it still has caller-lines
in the entries of the functions it calls.
</p>
<p>If the caller is part of a recursion cycle, the cycle number is
printed between the name and the index number.
</p></dd>
</dl>

<p>If the identity of the callers of a function cannot be determined, a
dummy caller-line is printed which has &lsquo;<samp>&lt;spontaneous&gt;</samp>&rsquo; as the
&ldquo;caller&rsquo;s name&rdquo; and all other fields blank.  This can happen for
signal handlers.
</p>
<hr>
<span id="Subroutines"></span><div class="header">
<p>
Next: <a href="#Cycles" accesskey="n" rel="next">Cycles</a>, Previous: <a href="#Callers" accesskey="p" rel="prev">Callers</a>, Up: <a href="#Call-Graph" accesskey="u" rel="up">Call Graph</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Lines-for-a-Function_0027s-Subroutines"></span><h4 class="subsection">5.2.3 Lines for a Function&rsquo;s Subroutines</h4>

<p>A function&rsquo;s entry has a line for each of its subroutines&mdash;in other
words, a line for each other function that it called.  These lines&rsquo;
fields correspond to the fields of the primary line, but their meanings
are different because of the difference in context.
</p>
<p>For reference, we repeat two lines from the entry for the function
<code>main</code>, the primary line and a line for a subroutine, together
with the heading line that shows the names of the fields:
</p>
<div class="example">
<pre class="example">index  % time    self  children called     name
&hellip;
[2]    100.0    0.00    0.05       1         main [2]
                0.00    0.05       1/1           report [3]
</pre></div>

<p>Here are the meanings of the fields in the subroutine-line for <code>main</code>
calling <code>report</code>:
</p>
<dl compact="compact">
<dt><code>self</code></dt>
<dd><p>An estimate of the amount of time spent directly within <code>report</code>
when <code>report</code> was called from <code>main</code>.
</p>
</dd>
<dt><code>children</code></dt>
<dd><p>An estimate of the amount of time spent in subroutines of <code>report</code>
when <code>report</code> was called from <code>main</code>.
</p>
<p>The sum of the <code>self</code> and <code>children</code> fields is an estimate
of the total time spent in calls to <code>report</code> from <code>main</code>.
</p>
</dd>
<dt><code>called</code></dt>
<dd><p>Two numbers, the number of calls to <code>report</code> from <code>main</code>
followed by the total number of non-recursive calls to <code>report</code>.
This ratio is used to determine how much of <code>report</code>&rsquo;s <code>self</code>
and <code>children</code> time gets credited to <code>main</code>.
See <a href="#Assumptions">Estimating <code>children</code> Times</a>.
</p>
</dd>
<dt><code>name</code></dt>
<dd><p>The name of the subroutine of <code>main</code> to which this line applies,
followed by the subroutine&rsquo;s index number.
</p>
<p>If the caller is part of a recursion cycle, the cycle number is
printed between the name and the index number.
</p></dd>
</dl>

<hr>
<span id="Cycles"></span><div class="header">
<p>
Previous: <a href="#Subroutines" accesskey="p" rel="prev">Subroutines</a>, Up: <a href="#Call-Graph" accesskey="u" rel="up">Call Graph</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="How-Mutually-Recursive-Functions-Are-Described"></span><h4 class="subsection">5.2.4 How Mutually Recursive Functions Are Described</h4>
<span id="index-cycle"></span>
<span id="index-recursion-cycle"></span>

<p>The graph may be complicated by the presence of <em>cycles of
recursion</em> in the call graph.  A cycle exists if a function calls
another function that (directly or indirectly) calls (or appears to
call) the original function.  For example: if <code>a</code> calls <code>b</code>,
and <code>b</code> calls <code>a</code>, then <code>a</code> and <code>b</code> form a cycle.
</p>
<p>Whenever there are call paths both ways between a pair of functions, they
belong to the same cycle.  If <code>a</code> and <code>b</code> call each other and
<code>b</code> and <code>c</code> call each other, all three make one cycle.  Note that
even if <code>b</code> only calls <code>a</code> if it was not called from <code>a</code>,
<code>gprof</code> cannot determine this, so <code>a</code> and <code>b</code> are still
considered a cycle.
</p>
<p>The cycles are numbered with consecutive integers.  When a function
belongs to a cycle, each time the function name appears in the call graph
it is followed by &lsquo;<samp>&lt;cycle <var>number</var>&gt;</samp>&rsquo;.
</p>
<p>The reason cycles matter is that they make the time values in the call
graph paradoxical.  The &ldquo;time spent in children&rdquo; of <code>a</code> should
include the time spent in its subroutine <code>b</code> and in <code>b</code>&rsquo;s
subroutines&mdash;but one of <code>b</code>&rsquo;s subroutines is <code>a</code>!  How much of
<code>a</code>&rsquo;s time should be included in the children of <code>a</code>, when
<code>a</code> is indirectly recursive?
</p>
<p>The way <code>gprof</code> resolves this paradox is by creating a single entry
for the cycle as a whole.  The primary line of this entry describes the
total time spent directly in the functions of the cycle.  The
&ldquo;subroutines&rdquo; of the cycle are the individual functions of the cycle, and
all other functions that were called directly by them.  The &ldquo;callers&rdquo; of
the cycle are the functions, outside the cycle, that called functions in
the cycle.
</p>
<p>Here is an example portion of a call graph which shows a cycle containing
functions <code>a</code> and <code>b</code>.  The cycle was entered by a call to
<code>a</code> from <code>main</code>; both <code>a</code> and <code>b</code> called <code>c</code>.
</p>
<div class="example">
<pre class="example">index  % time    self  children called     name
----------------------------------------
                 1.77        0    1/1        main [2]
[3]     91.71    1.77        0    1+5    &lt;cycle 1 as a whole&gt; [3]
                 1.02        0    3          b &lt;cycle 1&gt; [4]
                 0.75        0    2          a &lt;cycle 1&gt; [5]
----------------------------------------
                                  3          a &lt;cycle 1&gt; [5]
[4]     52.85    1.02        0    0      b &lt;cycle 1&gt; [4]
                                  2          a &lt;cycle 1&gt; [5]
                    0        0    3/6        c [6]
----------------------------------------
                 1.77        0    1/1        main [2]
                                  2          b &lt;cycle 1&gt; [4]
[5]     38.86    0.75        0    1      a &lt;cycle 1&gt; [5]
                                  3          b &lt;cycle 1&gt; [4]
                    0        0    3/6        c [6]
----------------------------------------
</pre></div>

<p>(The entire call graph for this program contains in addition an entry for
<code>main</code>, which calls <code>a</code>, and an entry for <code>c</code>, with callers
<code>a</code> and <code>b</code>.)
</p>
<div class="example">
<pre class="example">index  % time    self  children called     name
                                             &lt;spontaneous&gt;
[1]    100.00       0     1.93    0      start [1]
                 0.16     1.77    1/1        main [2]
----------------------------------------
                 0.16     1.77    1/1        start [1]
[2]    100.00    0.16     1.77    1      main [2]
                 1.77        0    1/1        a &lt;cycle 1&gt; [5]
----------------------------------------
                 1.77        0    1/1        main [2]
[3]     91.71    1.77        0    1+5    &lt;cycle 1 as a whole&gt; [3]
                 1.02        0    3          b &lt;cycle 1&gt; [4]
                 0.75        0    2          a &lt;cycle 1&gt; [5]
                    0        0    6/6        c [6]
----------------------------------------
                                  3          a &lt;cycle 1&gt; [5]
[4]     52.85    1.02        0    0      b &lt;cycle 1&gt; [4]
                                  2          a &lt;cycle 1&gt; [5]
                    0        0    3/6        c [6]
----------------------------------------
                 1.77        0    1/1        main [2]
                                  2          b &lt;cycle 1&gt; [4]
[5]     38.86    0.75        0    1      a &lt;cycle 1&gt; [5]
                                  3          b &lt;cycle 1&gt; [4]
                    0        0    3/6        c [6]
----------------------------------------
                    0        0    3/6        b &lt;cycle 1&gt; [4]
                    0        0    3/6        a &lt;cycle 1&gt; [5]
[6]      0.00       0        0    6      c [6]
----------------------------------------
</pre></div>

<p>The <code>self</code> field of the cycle&rsquo;s primary line is the total time
spent in all the functions of the cycle.  It equals the sum of the
<code>self</code> fields for the individual functions in the cycle, found
in the entry in the subroutine lines for these functions.
</p>
<p>The <code>children</code> fields of the cycle&rsquo;s primary line and subroutine lines
count only subroutines outside the cycle.  Even though <code>a</code> calls
<code>b</code>, the time spent in those calls to <code>b</code> is not counted in
<code>a</code>&rsquo;s <code>children</code> time.  Thus, we do not encounter the problem of
what to do when the time in those calls to <code>b</code> includes indirect
recursive calls back to <code>a</code>.
</p>
<p>The <code>children</code> field of a caller-line in the cycle&rsquo;s entry estimates
the amount of time spent <em>in the whole cycle</em>, and its other
subroutines, on the times when that caller called a function in the cycle.
</p>
<p>The <code>called</code> field in the primary line for the cycle has two numbers:
first, the number of times functions in the cycle were called by functions
outside the cycle; second, the number of times they were called by
functions in the cycle (including times when a function in the cycle calls
itself).  This is a generalization of the usual split into non-recursive and
recursive calls.
</p>
<p>The <code>called</code> field of a subroutine-line for a cycle member in the
cycle&rsquo;s entry says how many time that function was called from functions in
the cycle.  The total of all these is the second number in the primary line&rsquo;s
<code>called</code> field.
</p>
<p>In the individual entry for a function in a cycle, the other functions in
the same cycle can appear as subroutines and as callers.  These lines show
how many times each function in the cycle called or was called from each other
function in the cycle.  The <code>self</code> and <code>children</code> fields in these
lines are blank because of the difficulty of defining meanings for them
when recursion is going on.
</p>
<hr>
<span id="Line_002dby_002dline"></span><div class="header">
<p>
Next: <a href="#Annotated-Source" accesskey="n" rel="next">Annotated Source</a>, Previous: <a href="#Call-Graph" accesskey="p" rel="prev">Call Graph</a>, Up: <a href="#Output" accesskey="u" rel="up">Output</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Line_002dby_002dline-Profiling"></span><h3 class="section">5.3 Line-by-line Profiling</h3>

<p><code>gprof</code>&rsquo;s &lsquo;<samp>-l</samp>&rsquo; option causes the program to perform
<em>line-by-line</em> profiling.  In this mode, histogram
samples are assigned not to functions, but to individual
lines of source code.  This only works with programs compiled with
older versions of the <code>gcc</code> compiler.  Newer versions of <code>gcc</code>
use a different program - <code>gcov</code> - to display line-by-line
profiling information.
</p>
<p>With the older versions of <code>gcc</code> the program usually has to be
compiled with a &lsquo;<samp>-g</samp>&rsquo; option, in addition to &lsquo;<samp>-pg</samp>&rsquo;, in order
to generate debugging symbols for tracking source code lines.
Note, in much older versions of <code>gcc</code> the program had to be
compiled with the &lsquo;<samp>-a</samp>&rsquo; command-line option as well.
</p>
<p>The flat profile is the most useful output table
in line-by-line mode.
The call graph isn&rsquo;t as useful as normal, since
the current version of <code>gprof</code> does not propagate
call graph arcs from source code lines to the enclosing function.
The call graph does, however, show each line of code
that called each function, along with a count.
</p>
<p>Here is a section of <code>gprof</code>&rsquo;s output, without line-by-line profiling.
Note that <code>ct_init</code> accounted for four histogram hits, and
13327 calls to <code>init_block</code>.
</p>
<div class="example">
<pre class="example">Flat profile:

Each sample counts as 0.01 seconds.
  %   cumulative   self              self     total
 time   seconds   seconds    calls  us/call  us/call  name
 30.77      0.13     0.04     6335     6.31     6.31  ct_init


		     Call graph (explanation follows)


granularity: each sample hit covers 4 byte(s) for 7.69% of 0.13 seconds

index % time    self  children    called     name

                0.00    0.00       1/13496       name_too_long
                0.00    0.00      40/13496       deflate
                0.00    0.00     128/13496       deflate_fast
                0.00    0.00   13327/13496       ct_init
[7]      0.0    0.00    0.00   13496         init_block

</pre></div>

<p>Now let&rsquo;s look at some of <code>gprof</code>&rsquo;s output from the same program run,
this time with line-by-line profiling enabled.  Note that <code>ct_init</code>&rsquo;s
four histogram hits are broken down into four lines of source code&mdash;one hit
occurred on each of lines 349, 351, 382 and 385.  In the call graph,
note how
<code>ct_init</code>&rsquo;s 13327 calls to <code>init_block</code> are broken down
into one call from line 396, 3071 calls from line 384, 3730 calls
from line 385, and 6525 calls from 387.
</p>
<div class="example">
<pre class="example">Flat profile:

Each sample counts as 0.01 seconds.
  %   cumulative   self
 time   seconds   seconds    calls  name
  7.69      0.10     0.01           ct_init (trees.c:349)
  7.69      0.11     0.01           ct_init (trees.c:351)
  7.69      0.12     0.01           ct_init (trees.c:382)
  7.69      0.13     0.01           ct_init (trees.c:385)


		     Call graph (explanation follows)


granularity: each sample hit covers 4 byte(s) for 7.69% of 0.13 seconds

  % time    self  children    called     name

            0.00    0.00       1/13496       name_too_long (gzip.c:1440)
            0.00    0.00       1/13496       deflate (deflate.c:763)
            0.00    0.00       1/13496       ct_init (trees.c:396)
            0.00    0.00       2/13496       deflate (deflate.c:727)
            0.00    0.00       4/13496       deflate (deflate.c:686)
            0.00    0.00       5/13496       deflate (deflate.c:675)
            0.00    0.00      12/13496       deflate (deflate.c:679)
            0.00    0.00      16/13496       deflate (deflate.c:730)
            0.00    0.00     128/13496       deflate_fast (deflate.c:654)
            0.00    0.00    3071/13496       ct_init (trees.c:384)
            0.00    0.00    3730/13496       ct_init (trees.c:385)
            0.00    0.00    6525/13496       ct_init (trees.c:387)
[6]  0.0    0.00    0.00   13496         init_block (trees.c:408)

</pre></div>


<hr>
<span id="Annotated-Source"></span><div class="header">
<p>
Previous: <a href="#Line_002dby_002dline" accesskey="p" rel="prev">Line-by-line</a>, Up: <a href="#Output" accesskey="u" rel="up">Output</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="The-Annotated-Source-Listing"></span><h3 class="section">5.4 The Annotated Source Listing</h3>

<p><code>gprof</code>&rsquo;s &lsquo;<samp>-A</samp>&rsquo; option triggers an annotated source listing,
which lists the program&rsquo;s source code, each function labeled with the
number of times it was called.  You may also need to specify the
&lsquo;<samp>-I</samp>&rsquo; option, if <code>gprof</code> can&rsquo;t find the source code files.
</p>
<p>With older versions of <code>gcc</code> compiling with &lsquo;<samp>gcc &hellip; -g
-pg -a</samp>&rsquo; augments your program with basic-block counting code, in
addition to function counting code.  This enables <code>gprof</code> to
determine how many times each line of code was executed.  With newer
versions of <code>gcc</code> support for displaying basic-block counts is
provided by the <code>gcov</code> program.
</p>
<p>For example, consider the following function, taken from gzip,
with line numbers added:
</p>
<div class="example">
<pre class="example"> 1 ulg updcrc(s, n)
 2     uch *s;
 3     unsigned n;
 4 {
 5     register ulg c;
 6
 7     static ulg crc = (ulg)0xffffffffL;
 8
 9     if (s == NULL) {
10         c = 0xffffffffL;
11     } else {
12         c = crc;
13         if (n) do {
14             c = crc_32_tab[...];
15         } while (--n);
16     }
17     crc = c;
18     return c ^ 0xffffffffL;
19 }

</pre></div>

<p><code>updcrc</code> has at least five basic-blocks.
One is the function itself.  The
<code>if</code> statement on line 9 generates two more basic-blocks, one
for each branch of the <code>if</code>.  A fourth basic-block results from
the <code>if</code> on line 13, and the contents of the <code>do</code> loop form
the fifth basic-block.  The compiler may also generate additional
basic-blocks to handle various special cases.
</p>
<p>A program augmented for basic-block counting can be analyzed with
&lsquo;<samp>gprof -l -A</samp>&rsquo;.
The &lsquo;<samp>-x</samp>&rsquo; option is also helpful,
to ensure that each line of code is labeled at least once.
Here is <code>updcrc</code>&rsquo;s
annotated source listing for a sample <code>gzip</code> run:
</p>
<div class="example">
<pre class="example">                ulg updcrc(s, n)
                    uch *s;
                    unsigned n;
            2 -&gt;{
                    register ulg c;

                    static ulg crc = (ulg)0xffffffffL;

            2 -&gt;    if (s == NULL) {
            1 -&gt;        c = 0xffffffffL;
            1 -&gt;    } else {
            1 -&gt;        c = crc;
            1 -&gt;        if (n) do {
        26312 -&gt;            c = crc_32_tab[...];
26312,1,26311 -&gt;        } while (--n);
                    }
            2 -&gt;    crc = c;
            2 -&gt;    return c ^ 0xffffffffL;
            2 -&gt;}
</pre></div>

<p>In this example, the function was called twice, passing once through
each branch of the <code>if</code> statement.  The body of the <code>do</code>
loop was executed a total of 26312 times.  Note how the <code>while</code>
statement is annotated.  It began execution 26312 times, once for
each iteration through the loop.  One of those times (the last time)
it exited, while it branched back to the beginning of the loop 26311 times.
</p>
<hr>
<span id="Inaccuracy"></span><div class="header">
<p>
Next: <a href="#How-do-I_003f" accesskey="n" rel="next">How do I?</a>, Previous: <a href="#Output" accesskey="p" rel="prev">Output</a>, Up: <a href="#Top" accesskey="u" rel="up">Top</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Inaccuracy-of-gprof-Output"></span><h2 class="chapter">6 Inaccuracy of <code>gprof</code> Output</h2>

<table class="menu" border="0" cellspacing="0">
<tr><td align="left" valign="top">&bull; <a href="#Sampling-Error" accesskey="1">Sampling Error</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Statistical margins of error
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Assumptions" accesskey="2">Assumptions</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Estimating children times
</td></tr>
</table>

<hr>
<span id="Sampling-Error"></span><div class="header">
<p>
Next: <a href="#Assumptions" accesskey="n" rel="next">Assumptions</a>, Up: <a href="#Inaccuracy" accesskey="u" rel="up">Inaccuracy</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Statistical-Sampling-Error"></span><h3 class="section">6.1 Statistical Sampling Error</h3>

<p>The run-time figures that <code>gprof</code> gives you are based on a sampling
process, so they are subject to statistical inaccuracy.  If a function runs
only a small amount of time, so that on the average the sampling process
ought to catch that function in the act only once, there is a pretty good
chance it will actually find that function zero times, or twice.
</p>
<p>By contrast, the number-of-calls and basic-block figures are derived
by counting, not sampling.  They are completely accurate and will not
vary from run to run if your program is deterministic and single
threaded.  In multi-threaded applications, or single threaded
applications that link with multi-threaded libraries, the counts are
only deterministic if the counting function is thread-safe.  (Note:
beware that the mcount counting function in glibc is <em>not</em>
thread-safe).  See <a href="#Implementation">Implementation of Profiling</a>.
</p>
<p>The <em>sampling period</em> that is printed at the beginning of the flat
profile says how often samples are taken.  The rule of thumb is that a
run-time figure is accurate if it is considerably bigger than the sampling
period.
</p>
<p>The actual amount of error can be predicted.
For <var>n</var> samples, the <em>expected</em> error
is the square-root of <var>n</var>.  For example,
if the sampling period is 0.01 seconds and <code>foo</code>&rsquo;s run-time is 1 second,
<var>n</var> is 100 samples (1 second/0.01 seconds), sqrt(<var>n</var>) is 10 samples, so
the expected error in <code>foo</code>&rsquo;s run-time is 0.1 seconds (10*0.01 seconds),
or ten percent of the observed value.
Again, if the sampling period is 0.01 seconds and <code>bar</code>&rsquo;s run-time is
100 seconds, <var>n</var> is 10000 samples, sqrt(<var>n</var>) is 100 samples, so
the expected error in <code>bar</code>&rsquo;s run-time is 1 second,
or one percent of the observed value.
It is likely to
vary this much <em>on the average</em> from one profiling run to the next.
(<em>Sometimes</em> it will vary more.)
</p>
<p>This does not mean that a small run-time figure is devoid of information.
If the program&rsquo;s <em>total</em> run-time is large, a small run-time for one
function does tell you that that function used an insignificant fraction of
the whole program&rsquo;s time.  Usually this means it is not worth optimizing.
</p>
<p>One way to get more accuracy is to give your program more (but similar)
input data so it will take longer.  Another way is to combine the data from
several runs, using the &lsquo;<samp>-s</samp>&rsquo; option of <code>gprof</code>.  Here is how:
</p>
<ol>
<li> Run your program once.

</li><li> Issue the command &lsquo;<samp>mv gmon.out gmon.sum</samp>&rsquo;.

</li><li> Run your program again, the same as before.

</li><li> Merge the new data in <samp>gmon.out</samp> into <samp>gmon.sum</samp> with this command:

<div class="example">
<pre class="example">gprof -s <var>executable-file</var> gmon.out gmon.sum
</pre></div>

</li><li> Repeat the last two steps as often as you wish.

</li><li> Analyze the cumulative data using this command:

<div class="example">
<pre class="example">gprof <var>executable-file</var> gmon.sum &gt; <var>output-file</var>
</pre></div>
</li></ol>

<hr>
<span id="Assumptions"></span><div class="header">
<p>
Previous: <a href="#Sampling-Error" accesskey="p" rel="prev">Sampling Error</a>, Up: <a href="#Inaccuracy" accesskey="u" rel="up">Inaccuracy</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Estimating-children-Times"></span><h3 class="section">6.2 Estimating <code>children</code> Times</h3>

<p>Some of the figures in the call graph are estimates&mdash;for example, the
<code>children</code> time values and all the time figures in caller and
subroutine lines.
</p>
<p>There is no direct information about these measurements in the profile
data itself.  Instead, <code>gprof</code> estimates them by making an assumption
about your program that might or might not be true.
</p>
<p>The assumption made is that the average time spent in each call to any
function <code>foo</code> is not correlated with who called <code>foo</code>.  If
<code>foo</code> used 5 seconds in all, and 2/5 of the calls to <code>foo</code> came
from <code>a</code>, then <code>foo</code> contributes 2 seconds to <code>a</code>&rsquo;s
<code>children</code> time, by assumption.
</p>
<p>This assumption is usually true enough, but for some programs it is far
from true.  Suppose that <code>foo</code> returns very quickly when its argument
is zero; suppose that <code>a</code> always passes zero as an argument, while
other callers of <code>foo</code> pass other arguments.  In this program, all the
time spent in <code>foo</code> is in the calls from callers other than <code>a</code>.
But <code>gprof</code> has no way of knowing this; it will blindly and
incorrectly charge 2 seconds of time in <code>foo</code> to the children of
<code>a</code>.
</p>
<p>We hope some day to put more complete data into <samp>gmon.out</samp>, so that
this assumption is no longer needed, if we can figure out how.  For the
novice, the estimated figures are usually more useful than misleading.
</p>
<hr>
<span id="How-do-I_003f"></span><div class="header">
<p>
Next: <a href="#Incompatibilities" accesskey="n" rel="next">Incompatibilities</a>, Previous: <a href="#Inaccuracy" accesskey="p" rel="prev">Inaccuracy</a>, Up: <a href="#Top" accesskey="u" rel="up">Top</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Answers-to-Common-Questions"></span><h2 class="chapter">7 Answers to Common Questions</h2>

<dl compact="compact">
<dt>How can I get more exact information about hot spots in my program?</dt>
<dd>
<p>Looking at the per-line call counts only tells part of the story.
Because <code>gprof</code> can only report call times and counts by function,
the best way to get finer-grained information on where the program
is spending its time is to re-factor large functions into sequences
of calls to smaller ones.  Beware however that this can introduce
artificial hot spots since compiling with &lsquo;<samp>-pg</samp>&rsquo; adds a significant
overhead to function calls.  An alternative solution is to use a
non-intrusive profiler, e.g. oprofile.
</p>
</dd>
<dt>How do I find which lines in my program were executed the most times?</dt>
<dd>
<p>Use the <code>gcov</code> program.
</p>
</dd>
<dt>How do I find which lines in my program called a particular function?</dt>
<dd>
<p>Use &lsquo;<samp>gprof -l</samp>&rsquo; and lookup the function in the call graph.
The callers will be broken down by function and line number.
</p>
</dd>
<dt>How do I analyze a program that runs for less than a second?</dt>
<dd>
<p>Try using a shell script like this one:
</p>
<div class="example">
<pre class="example">for i in `seq 1 100`; do
  fastprog
  mv gmon.out gmon.out.$i
done

gprof -s fastprog gmon.out.*

gprof fastprog gmon.sum
</pre></div>

<p>If your program is completely deterministic, all the call counts
will be simple multiples of 100 (i.e., a function called once in
each run will appear with a call count of 100).
</p>
</dd>
</dl>

<hr>
<span id="Incompatibilities"></span><div class="header">
<p>
Next: <a href="#Details" accesskey="n" rel="next">Details</a>, Previous: <a href="#How-do-I_003f" accesskey="p" rel="prev">How do I?</a>, Up: <a href="#Top" accesskey="u" rel="up">Top</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Incompatibilities-with-Unix-gprof"></span><h2 class="chapter">8 Incompatibilities with Unix <code>gprof</code></h2>

<p><small>GNU</small> <code>gprof</code> and Berkeley Unix <code>gprof</code> use the same data
file <samp>gmon.out</samp>, and provide essentially the same information.  But
there are a few differences.
</p>
<ul>
<li> <small>GNU</small> <code>gprof</code> uses a new, generalized file format with support
for basic-block execution counts and non-realtime histograms.  A magic
cookie and version number allows <code>gprof</code> to easily identify
new style files.  Old BSD-style files can still be read.
See <a href="#File-Format">Profiling Data File Format</a>.

</li><li> For a recursive function, Unix <code>gprof</code> lists the function as a
parent and as a child, with a <code>calls</code> field that lists the number
of recursive calls.  <small>GNU</small> <code>gprof</code> omits these lines and puts
the number of recursive calls in the primary line.

</li><li> When a function is suppressed from the call graph with &lsquo;<samp>-e</samp>&rsquo;, <small>GNU</small>
<code>gprof</code> still lists it as a subroutine of functions that call it.

</li><li> <small>GNU</small> <code>gprof</code> accepts the &lsquo;<samp>-k</samp>&rsquo; with its argument
in the form &lsquo;<samp>from/to</samp>&rsquo;, instead of &lsquo;<samp>from to</samp>&rsquo;.

</li><li> In the annotated source listing,
if there are multiple basic blocks on the same line,
<small>GNU</small> <code>gprof</code> prints all of their counts, separated by commas.


</li><li> The blurbs, field widths, and output formats are different.  <small>GNU</small>
<code>gprof</code> prints blurbs after the tables, so that you can see the
tables without skipping the blurbs.
</li></ul>

<hr>
<span id="Details"></span><div class="header">
<p>
Next: <a href="#GNU-Free-Documentation-License" accesskey="n" rel="next">GNU Free Documentation License</a>, Previous: <a href="#Incompatibilities" accesskey="p" rel="prev">Incompatibilities</a>, Up: <a href="#Top" accesskey="u" rel="up">Top</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Details-of-Profiling"></span><h2 class="chapter">9 Details of Profiling</h2>

<table class="menu" border="0" cellspacing="0">
<tr><td align="left" valign="top">&bull; <a href="#Implementation" accesskey="1">Implementation</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">How a program collects profiling information
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#File-Format" accesskey="2">File Format</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Format of &lsquo;<samp>gmon.out</samp>&rsquo; files
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Internals" accesskey="3">Internals</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top"><code>gprof</code>&rsquo;s internal operation
</td></tr>
<tr><td align="left" valign="top">&bull; <a href="#Debugging" accesskey="4">Debugging</a></td><td>&nbsp;&nbsp;</td><td align="left" valign="top">Using <code>gprof</code>&rsquo;s &lsquo;<samp>-d</samp>&rsquo; option
</td></tr>
</table>

<hr>
<span id="Implementation"></span><div class="header">
<p>
Next: <a href="#File-Format" accesskey="n" rel="next">File Format</a>, Up: <a href="#Details" accesskey="u" rel="up">Details</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Implementation-of-Profiling"></span><h3 class="section">9.1 Implementation of Profiling</h3>

<p>Profiling works by changing how every function in your program is compiled
so that when it is called, it will stash away some information about where
it was called from.  From this, the profiler can figure out what function
called it, and can count how many times it was called.  This change is made
by the compiler when your program is compiled with the &lsquo;<samp>-pg</samp>&rsquo; option,
which causes every function to call <code>mcount</code>
(or <code>_mcount</code>, or <code>__mcount</code>, depending on the OS and compiler)
as one of its first operations.
</p>
<p>The <code>mcount</code> routine, included in the profiling library,
is responsible for recording in an in-memory call graph table
both its parent routine (the child) and its parent&rsquo;s parent.  This is
typically done by examining the stack frame to find both
the address of the child, and the return address in the original parent.
Since this is a very machine-dependent operation, <code>mcount</code>
itself is typically a short assembly-language stub routine
that extracts the required
information, and then calls <code>__mcount_internal</code>
(a normal C function) with two arguments&mdash;<code>frompc</code> and <code>selfpc</code>.
<code>__mcount_internal</code> is responsible for maintaining
the in-memory call graph, which records <code>frompc</code>, <code>selfpc</code>,
and the number of times each of these call arcs was traversed.
</p>
<p>GCC Version 2 provides a magical function (<code>__builtin_return_address</code>),
which allows a generic <code>mcount</code> function to extract the
required information from the stack frame.  However, on some
architectures, most notably the SPARC, using this builtin can be
very computationally expensive, and an assembly language version
of <code>mcount</code> is used for performance reasons.
</p>
<p>Number-of-calls information for library routines is collected by using a
special version of the C library.  The programs in it are the same as in
the usual C library, but they were compiled with &lsquo;<samp>-pg</samp>&rsquo;.  If you
link your program with &lsquo;<samp>gcc &hellip; -pg</samp>&rsquo;, it automatically uses the
profiling version of the library.
</p>
<p>Profiling also involves watching your program as it runs, and keeping a
histogram of where the program counter happens to be every now and then.
Typically the program counter is looked at around 100 times per second of
run time, but the exact frequency may vary from system to system.
</p>
<p>This is done is one of two ways.  Most UNIX-like operating systems
provide a <code>profil()</code> system call, which registers a memory
array with the kernel, along with a scale
factor that determines how the program&rsquo;s address space maps
into the array.
Typical scaling values cause every 2 to 8 bytes of address space
to map into a single array slot.
On every tick of the system clock
(assuming the profiled program is running), the value of the
program counter is examined and the corresponding slot in
the memory array is incremented.  Since this is done in the kernel,
which had to interrupt the process anyway to handle the clock
interrupt, very little additional system overhead is required.
</p>
<p>However, some operating systems, most notably Linux 2.0 (and earlier),
do not provide a <code>profil()</code> system call.  On such a system,
arrangements are made for the kernel to periodically deliver
a signal to the process (typically via <code>setitimer()</code>),
which then performs the same operation of examining the
program counter and incrementing a slot in the memory array.
Since this method requires a signal to be delivered to
user space every time a sample is taken, it uses considerably
more overhead than kernel-based profiling.  Also, due to the
added delay required to deliver the signal, this method is
less accurate as well.
</p>
<p>A special startup routine allocates memory for the histogram and
either calls <code>profil()</code> or sets up
a clock signal handler.
This routine (<code>monstartup</code>) can be invoked in several ways.
On Linux systems, a special profiling startup file <code>gcrt0.o</code>,
which invokes <code>monstartup</code> before <code>main</code>,
is used instead of the default <code>crt0.o</code>.
Use of this special startup file is one of the effects
of using &lsquo;<samp>gcc &hellip; -pg</samp>&rsquo; to link.
On SPARC systems, no special startup files are used.
Rather, the <code>mcount</code> routine, when it is invoked for
the first time (typically when <code>main</code> is called),
calls <code>monstartup</code>.
</p>
<p>If the compiler&rsquo;s &lsquo;<samp>-a</samp>&rsquo; option was used, basic-block counting
is also enabled.  Each object file is then compiled with a static array
of counts, initially zero.
In the executable code, every time a new basic-block begins
(i.e., when an <code>if</code> statement appears), an extra instruction
is inserted to increment the corresponding count in the array.
At compile time, a paired array was constructed that recorded
the starting address of each basic-block.  Taken together,
the two arrays record the starting address of every basic-block,
along with the number of times it was executed.
</p>
<p>The profiling library also includes a function (<code>mcleanup</code>) which is
typically registered using <code>atexit()</code> to be called as the
program exits, and is responsible for writing the file <samp>gmon.out</samp>.
Profiling is turned off, various headers are output, and the histogram
is written, followed by the call-graph arcs and the basic-block counts.
</p>
<p>The output from <code>gprof</code> gives no indication of parts of your program that
are limited by I/O or swapping bandwidth.  This is because samples of the
program counter are taken at fixed intervals of the program&rsquo;s run time.
Therefore, the
time measurements in <code>gprof</code> output say nothing about time that your
program was not running.  For example, a part of the program that creates
so much data that it cannot all fit in physical memory at once may run very
slowly due to thrashing, but <code>gprof</code> will say it uses little time.  On
the other hand, sampling by run time has the advantage that the amount of
load due to other users won&rsquo;t directly affect the output you get.
</p>
<hr>
<span id="File-Format"></span><div class="header">
<p>
Next: <a href="#Internals" accesskey="n" rel="next">Internals</a>, Previous: <a href="#Implementation" accesskey="p" rel="prev">Implementation</a>, Up: <a href="#Details" accesskey="u" rel="up">Details</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Profiling-Data-File-Format"></span><h3 class="section">9.2 Profiling Data File Format</h3>

<p>The old BSD-derived file format used for profile data does not contain a
magic cookie that allows one to check whether a data file really is a
<code>gprof</code> file.  Furthermore, it does not provide a version number, thus
rendering changes to the file format almost impossible.  <small>GNU</small> <code>gprof</code>
uses a new file format that provides these features.  For backward
compatibility, <small>GNU</small> <code>gprof</code> continues to support the old BSD-derived
format, but not all features are supported with it.  For example,
basic-block execution counts cannot be accommodated by the old file
format.
</p>
<p>The new file format is defined in header file <samp>gmon_out.h</samp>.  It
consists of a header containing the magic cookie and a version number,
as well as some spare bytes available for future extensions.  All data
in a profile data file is in the native format of the target for which
the profile was collected.  <small>GNU</small> <code>gprof</code> adapts automatically
to the byte-order in use.
</p>
<p>In the new file format, the header is followed by a sequence of
records.  Currently, there are three different record types: histogram
records, call-graph arc records, and basic-block execution count
records.  Each file can contain any number of each record type.  When
reading a file, <small>GNU</small> <code>gprof</code> will ensure records of the same type are
compatible with each other and compute the union of all records.  For
example, for basic-block execution counts, the union is simply the sum
of all execution counts for each basic-block.
</p>
<span id="Histogram-Records"></span><h4 class="subsection">9.2.1 Histogram Records</h4>

<p>Histogram records consist of a header that is followed by an array of
bins.  The header contains the text-segment range that the histogram
spans, the size of the histogram in bytes (unlike in the old BSD
format, this does not include the size of the header), the rate of the
profiling clock, and the physical dimension that the bin counts
represent after being scaled by the profiling clock rate.  The
physical dimension is specified in two parts: a long name of up to 15
characters and a single character abbreviation.  For example, a
histogram representing real-time would specify the long name as
&ldquo;seconds&rdquo; and the abbreviation as &ldquo;s&rdquo;.  This feature is useful for
architectures that support performance monitor hardware (which,
fortunately, is becoming increasingly common).  For example, under DEC
OSF/1, the &ldquo;uprofile&rdquo; command can be used to produce a histogram of,
say, instruction cache misses.  In this case, the dimension in the
histogram header could be set to &ldquo;i-cache misses&rdquo; and the abbreviation
could be set to &ldquo;1&rdquo; (because it is simply a count, not a physical
dimension).  Also, the profiling rate would have to be set to 1 in
this case.
</p>
<p>Histogram bins are 16-bit numbers and each bin represent an equal
amount of text-space.  For example, if the text-segment is one
thousand bytes long and if there are ten bins in the histogram, each
bin represents one hundred bytes.
</p>

<span id="Call_002dGraph-Records"></span><h4 class="subsection">9.2.2 Call-Graph Records</h4>

<p>Call-graph records have a format that is identical to the one used in
the BSD-derived file format.  It consists of an arc in the call graph
and a count indicating the number of times the arc was traversed
during program execution.  Arcs are specified by a pair of addresses:
the first must be within caller&rsquo;s function and the second must be
within the callee&rsquo;s function.  When performing profiling at the
function level, these addresses can point anywhere within the
respective function.  However, when profiling at the line-level, it is
better if the addresses are as close to the call-site/entry-point as
possible.  This will ensure that the line-level call-graph is able to
identify exactly which line of source code performed calls to a
function.
</p>
<span id="Basic_002dBlock-Execution-Count-Records"></span><h4 class="subsection">9.2.3 Basic-Block Execution Count Records</h4>

<p>Basic-block execution count records consist of a header followed by a
sequence of address/count pairs.  The header simply specifies the
length of the sequence.  In an address/count pair, the address
identifies a basic-block and the count specifies the number of times
that basic-block was executed.  Any address within the basic-address can
be used.
</p>
<hr>
<span id="Internals"></span><div class="header">
<p>
Next: <a href="#Debugging" accesskey="n" rel="next">Debugging</a>, Previous: <a href="#File-Format" accesskey="p" rel="prev">File Format</a>, Up: <a href="#Details" accesskey="u" rel="up">Details</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="gprof_0027s-Internal-Operation"></span><h3 class="section">9.3 <code>gprof</code>&rsquo;s Internal Operation</h3>

<p>Like most programs, <code>gprof</code> begins by processing its options.
During this stage, it may building its symspec list
(<code>sym_ids.c:sym_id_add</code>), if
options are specified which use symspecs.
<code>gprof</code> maintains a single linked list of symspecs,
which will eventually get turned into 12 symbol tables,
organized into six include/exclude pairs&mdash;one
pair each for the flat profile (INCL_FLAT/EXCL_FLAT),
the call graph arcs (INCL_ARCS/EXCL_ARCS),
printing in the call graph (INCL_GRAPH/EXCL_GRAPH),
timing propagation in the call graph (INCL_TIME/EXCL_TIME),
the annotated source listing (INCL_ANNO/EXCL_ANNO),
and the execution count listing (INCL_EXEC/EXCL_EXEC).
</p>
<p>After option processing, <code>gprof</code> finishes
building the symspec list by adding all the symspecs in
<code>default_excluded_list</code> to the exclude lists
EXCL_TIME and EXCL_GRAPH, and if line-by-line profiling is specified,
EXCL_FLAT as well.
These default excludes are not added to EXCL_ANNO, EXCL_ARCS, and EXCL_EXEC.
</p>
<p>Next, the BFD library is called to open the object file,
verify that it is an object file,
and read its symbol table (<code>core.c:core_init</code>),
using <code>bfd_canonicalize_symtab</code> after mallocing
an appropriately sized array of symbols.  At this point,
function mappings are read (if the &lsquo;<samp>--file-ordering</samp>&rsquo; option
has been specified), and the core text space is read into
memory (if the &lsquo;<samp>-c</samp>&rsquo; option was given).
</p>
<p><code>gprof</code>&rsquo;s own symbol table, an array of Sym structures,
is now built.
This is done in one of two ways, by one of two routines, depending
on whether line-by-line profiling (&lsquo;<samp>-l</samp>&rsquo; option) has been
enabled.
For normal profiling, the BFD canonical symbol table is scanned.
For line-by-line profiling, every
text space address is examined, and a new symbol table entry
gets created every time the line number changes.
In either case, two passes are made through the symbol
table&mdash;one to count the size of the symbol table required,
and the other to actually read the symbols.  In between the
two passes, a single array of type <code>Sym</code> is created of
the appropriate length.
Finally, <code>symtab.c:symtab_finalize</code>
is called to sort the symbol table and remove duplicate entries
(entries with the same memory address).
</p>
<p>The symbol table must be a contiguous array for two reasons.
First, the <code>qsort</code> library function (which sorts an array)
will be used to sort the symbol table.
Also, the symbol lookup routine (<code>symtab.c:sym_lookup</code>),
which finds symbols
based on memory address, uses a binary search algorithm
which requires the symbol table to be a sorted array.
Function symbols are indicated with an <code>is_func</code> flag.
Line number symbols have no special flags set.
Additionally, a symbol can have an <code>is_static</code> flag
to indicate that it is a local symbol.
</p>
<p>With the symbol table read, the symspecs can now be translated
into Syms (<code>sym_ids.c:sym_id_parse</code>).  Remember that a single
symspec can match multiple symbols.
An array of symbol tables
(<code>syms</code>) is created, each entry of which is a symbol table
of Syms to be included or excluded from a particular listing.
The master symbol table and the symspecs are examined by nested
loops, and every symbol that matches a symspec is inserted
into the appropriate syms table.  This is done twice, once to
count the size of each required symbol table, and again to build
the tables, which have been malloced between passes.
From now on, to determine whether a symbol is on an include
or exclude symspec list, <code>gprof</code> simply uses its
standard symbol lookup routine on the appropriate table
in the <code>syms</code> array.
</p>
<p>Now the profile data file(s) themselves are read
(<code>gmon_io.c:gmon_out_read</code>),
first by checking for a new-style &lsquo;<samp>gmon.out</samp>&rsquo; header,
then assuming this is an old-style BSD &lsquo;<samp>gmon.out</samp>&rsquo;
if the magic number test failed.
</p>
<p>New-style histogram records are read by <code>hist.c:hist_read_rec</code>.
For the first histogram record, allocate a memory array to hold
all the bins, and read them in.
When multiple profile data files (or files with multiple histogram
records) are read, the memory ranges of each pair of histogram records
must be either equal, or non-overlapping.  For each pair of histogram
records, the resolution (memory region size divided by the number of
bins) must be the same.  The time unit must be the same for all
histogram records. If the above containts are met, all histograms
for the same memory range are merged.
</p>
<p>As each call graph record is read (<code>call_graph.c:cg_read_rec</code>),
the parent and child addresses
are matched to symbol table entries, and a call graph arc is
created by <code>cg_arcs.c:arc_add</code>, unless the arc fails a symspec
check against INCL_ARCS/EXCL_ARCS.  As each arc is added,
a linked list is maintained of the parent&rsquo;s child arcs, and of the child&rsquo;s
parent arcs.
Both the child&rsquo;s call count and the arc&rsquo;s call count are
incremented by the record&rsquo;s call count.
</p>
<p>Basic-block records are read (<code>basic_blocks.c:bb_read_rec</code>),
but only if line-by-line profiling has been selected.
Each basic-block address is matched to a corresponding line
symbol in the symbol table, and an entry made in the symbol&rsquo;s
bb_addr and bb_calls arrays.  Again, if multiple basic-block
records are present for the same address, the call counts
are cumulative.
</p>
<p>A gmon.sum file is dumped, if requested (<code>gmon_io.c:gmon_out_write</code>).
</p>
<p>If histograms were present in the data files, assign them to symbols
(<code>hist.c:hist_assign_samples</code>) by iterating over all the sample
bins and assigning them to symbols.  Since the symbol table
is sorted in order of ascending memory addresses, we can
simple follow along in the symbol table as we make our pass
over the sample bins.
This step includes a symspec check against INCL_FLAT/EXCL_FLAT.
Depending on the histogram
scale factor, a sample bin may span multiple symbols,
in which case a fraction of the sample count is allocated
to each symbol, proportional to the degree of overlap.
This effect is rare for normal profiling, but overlaps
are more common during line-by-line profiling, and can
cause each of two adjacent lines to be credited with half
a hit, for example.
</p>
<p>If call graph data is present, <code>cg_arcs.c:cg_assemble</code> is called.
First, if &lsquo;<samp>-c</samp>&rsquo; was specified, a machine-dependent
routine (<code>find_call</code>) scans through each symbol&rsquo;s machine code,
looking for subroutine call instructions, and adding them
to the call graph with a zero call count.
A topological sort is performed by depth-first numbering
all the symbols (<code>cg_dfn.c:cg_dfn</code>), so that
children are always numbered less than their parents,
then making a array of pointers into the symbol table and sorting it into
numerical order, which is reverse topological
order (children appear before parents).
Cycles are also detected at this point, all members
of which are assigned the same topological number.
Two passes are now made through this sorted array of symbol pointers.
The first pass, from end to beginning (parents to children),
computes the fraction of child time to propagate to each parent
and a print flag.
The print flag reflects symspec handling of INCL_GRAPH/EXCL_GRAPH,
with a parent&rsquo;s include or exclude (print or no print) property
being propagated to its children, unless they themselves explicitly appear
in INCL_GRAPH or EXCL_GRAPH.
A second pass, from beginning to end (children to parents) actually
propagates the timings along the call graph, subject
to a check against INCL_TIME/EXCL_TIME.
With the print flag, fractions, and timings now stored in the symbol
structures, the topological sort array is now discarded, and a
new array of pointers is assembled, this time sorted by propagated time.
</p>
<p>Finally, print the various outputs the user requested, which is now fairly
straightforward.  The call graph (<code>cg_print.c:cg_print</code>) and
flat profile (<code>hist.c:hist_print</code>) are regurgitations of values
already computed.  The annotated source listing
(<code>basic_blocks.c:print_annotated_source</code>) uses basic-block
information, if present, to label each line of code with call counts,
otherwise only the function call counts are presented.
</p>
<p>The function ordering code is marginally well documented
in the source code itself (<code>cg_print.c</code>).  Basically,
the functions with the most use and the most parents are
placed first, followed by other functions with the most use,
followed by lower use functions, followed by unused functions
at the end.
</p>
<hr>
<span id="Debugging"></span><div class="header">
<p>
Previous: <a href="#Internals" accesskey="p" rel="prev">Internals</a>, Up: <a href="#Details" accesskey="u" rel="up">Details</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="Debugging-gprof"></span><h3 class="section">9.4 Debugging <code>gprof</code></h3>

<p>If <code>gprof</code> was compiled with debugging enabled,
the &lsquo;<samp>-d</samp>&rsquo; option triggers debugging output
(to stdout) which can be helpful in understanding its operation.
The debugging number specified is interpreted as a sum of the following
options:
</p>
<dl compact="compact">
<dt>2 - Topological sort</dt>
<dd><p>Monitor depth-first numbering of symbols during call graph analysis
</p></dd>
<dt>4 - Cycles</dt>
<dd><p>Shows symbols as they are identified as cycle heads
</p></dd>
<dt>16 - Tallying</dt>
<dd><p>As the call graph arcs are read, show each arc and how
the total calls to each function are tallied
</p></dd>
<dt>32 - Call graph arc sorting</dt>
<dd><p>Details sorting individual parents/children within each call graph entry
</p></dd>
<dt>64 - Reading histogram and call graph records</dt>
<dd><p>Shows address ranges of histograms as they are read, and each
call graph arc
</p></dd>
<dt>128 - Symbol table</dt>
<dd><p>Reading, classifying, and sorting the symbol table from the object file.
For line-by-line profiling (&lsquo;<samp>-l</samp>&rsquo; option), also shows line numbers
being assigned to memory addresses.
</p></dd>
<dt>256 - Static call graph</dt>
<dd><p>Trace operation of &lsquo;<samp>-c</samp>&rsquo; option
</p></dd>
<dt>512 - Symbol table and arc table lookups</dt>
<dd><p>Detail operation of lookup routines
</p></dd>
<dt>1024 - Call graph propagation</dt>
<dd><p>Shows how function times are propagated along the call graph
</p></dd>
<dt>2048 - Basic-blocks</dt>
<dd><p>Shows basic-block records as they are read from profile data
(only meaningful with &lsquo;<samp>-l</samp>&rsquo; option)
</p></dd>
<dt>4096 - Symspecs</dt>
<dd><p>Shows symspec-to-symbol pattern matching operation
</p></dd>
<dt>8192 - Annotate source</dt>
<dd><p>Tracks operation of &lsquo;<samp>-A</samp>&rsquo; option
</p></dd>
</dl>

<hr>
<span id="GNU-Free-Documentation-License"></span><div class="header">
<p>
Previous: <a href="#Details" accesskey="p" rel="prev">Details</a>, Up: <a href="#Top" accesskey="u" rel="up">Top</a> &nbsp; [<a href="#SEC_Contents" title="Table of contents" rel="contents">Contents</a>]</p>
</div>
<span id="GNU-Free-Documentation-License-1"></span><h2 class="appendix">Appendix A GNU Free Documentation License</h2>
<div align="center">Version 1.3, 3 November 2008
</div>

<div class="display">
<pre class="display">Copyright &copy; 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc.
<a href="http://fsf.org/">http://fsf.org/</a>

Everyone is permitted to copy and distribute verbatim copies
of this license document, but changing it is not allowed.
</pre></div>

<ol start="0">
<li> PREAMBLE

<p>The purpose of this License is to make a manual, textbook, or other
functional and useful document <em>free</em> in the sense of freedom: to
assure everyone the effective freedom to copy and redistribute it,
with or without modifying it, either commercially or noncommercially.
Secondarily, this License preserves for the author and publisher a way
to get credit for their work, while not being considered responsible
for modifications made by others.
</p>
<p>This License is a kind of &ldquo;copyleft&rdquo;, which means that derivative
works of the document must themselves be free in the same sense.  It
complements the GNU General Public License, which is a copyleft
license designed for free software.
</p>
<p>We have designed this License in order to use it for manuals for free
software, because free software needs free documentation: a free
program should come with manuals providing the same freedoms that the
software does.  But this License is not limited to software manuals;
it can be used for any textual work, regardless of subject matter or
whether it is published as a printed book.  We recommend this License
principally for works whose purpose is instruction or reference.
</p>
</li><li> APPLICABILITY AND DEFINITIONS

<p>This License applies to any manual or other work, in any medium, that
contains a notice placed by the copyright holder saying it can be
distributed under the terms of this License.  Such a notice grants a
world-wide, royalty-free license, unlimited in duration, to use that
work under the conditions stated herein.  The &ldquo;Document&rdquo;, below,
refers to any such manual or work.  Any member of the public is a
licensee, and is addressed as &ldquo;you&rdquo;.  You accept the license if you
copy, modify or distribute the work in a way requiring permission
under copyright law.
</p>
<p>A &ldquo;Modified Version&rdquo; of the Document means any work containing the
Document or a portion of it, either copied verbatim, or with
modifications and/or translated into another language.
</p>
<p>A &ldquo;Secondary Section&rdquo; is a named appendix or a front-matter section
of the Document that deals exclusively with the relationship of the
publishers or authors of the Document to the Document&rsquo;s overall
subject (or to related matters) and contains nothing that could fall
directly within that overall subject.  (Thus, if the Document is in
part a textbook of mathematics, a Secondary Section may not explain
any mathematics.)  The relationship could be a matter of historical
connection with the subject or with related matters, or of legal,
commercial, philosophical, ethical or political position regarding
them.
</p>
<p>The &ldquo;Invariant Sections&rdquo; are certain Secondary Sections whose titles
are designated, as being those of Invariant Sections, in the notice
that says that the Document is released under this License.  If a
section does not fit the above definition of Secondary then it is not
allowed to be designated as Invariant.  The Document may contain zero
Invariant Sections.  If the Document does not identify any Invariant
Sections then there are none.
</p>
<p>The &ldquo;Cover Texts&rdquo; are certain short passages of text that are listed,
as Front-Cover Texts or Back-Cover Texts, in the notice that says that
the Document is released under this License.  A Front-Cover Text may
be at most 5 words, and a Back-Cover Text may be at most 25 words.
</p>
<p>A &ldquo;Transparent&rdquo; copy of the Document means a machine-readable copy,
represented in a format whose specification is available to the
general public, that is suitable for revising the document
straightforwardly with generic text editors or (for images composed of
pixels) generic paint programs or (for drawings) some widely available
drawing editor, and that is suitable for input to text formatters or
for automatic translation to a variety of formats suitable for input
to text formatters.  A copy made in an otherwise Transparent file
format whose markup, or absence of markup, has been arranged to thwart
or discourage subsequent modification by readers is not Transparent.
An image format is not Transparent if used for any substantial amount
of text.  A copy that is not &ldquo;Transparent&rdquo; is called &ldquo;Opaque&rdquo;.
</p>
<p>Examples of suitable formats for Transparent copies include plain
<small>ASCII</small> without markup, Texinfo input format, LaTeX input
format, <acronym>SGML</acronym> or <acronym>XML</acronym> using a publicly available
<acronym>DTD</acronym>, and standard-conforming simple <acronym>HTML</acronym>,
PostScript or <acronym>PDF</acronym> designed for human modification.  Examples
of transparent image formats include <acronym>PNG</acronym>, <acronym>XCF</acronym> and
<acronym>JPG</acronym>.  Opaque formats include proprietary formats that can be
read and edited only by proprietary word processors, <acronym>SGML</acronym> or
<acronym>XML</acronym> for which the <acronym>DTD</acronym> and/or processing tools are
not generally available, and the machine-generated <acronym>HTML</acronym>,
PostScript or <acronym>PDF</acronym> produced by some word processors for
output purposes only.
</p>
<p>The &ldquo;Title Page&rdquo; means, for a printed book, the title page itself,
plus such following pages as are needed to hold, legibly, the material
this License requires to appear in the title page.  For works in
formats which do not have any title page as such, &ldquo;Title Page&rdquo; means
the text near the most prominent appearance of the work&rsquo;s title,
preceding the beginning of the body of the text.
</p>
<p>The &ldquo;publisher&rdquo; means any person or entity that distributes copies
of the Document to the public.
</p>
<p>A section &ldquo;Entitled XYZ&rdquo; means a named subunit of the Document whose
title either is precisely XYZ or contains XYZ in parentheses following
text that translates XYZ in another language.  (Here XYZ stands for a
specific section name mentioned below, such as &ldquo;Acknowledgements&rdquo;,
&ldquo;Dedications&rdquo;, &ldquo;Endorsements&rdquo;, or &ldquo;History&rdquo;.)  To &ldquo;Preserve the Title&rdquo;
of such a section when you modify the Document means that it remains a
section &ldquo;Entitled XYZ&rdquo; according to this definition.
</p>
<p>The Document may include Warranty Disclaimers next to the notice which
states that this License applies to the Document.  These Warranty
Disclaimers are considered to be included by reference in this
License, but only as regards disclaiming warranties: any other
implication that these Warranty Disclaimers may have is void and has
no effect on the meaning of this License.
</p>
</li><li> VERBATIM COPYING

<p>You may copy and distribute the Document in any medium, either
commercially or noncommercially, provided that this License, the
copyright notices, and the license notice saying this License applies
to the Document are reproduced in all copies, and that you add no other
conditions whatsoever to those of this License.  You may not use
technical measures to obstruct or control the reading or further
copying of the copies you make or distribute.  However, you may accept
compensation in exchange for copies.  If you distribute a large enough
number of copies you must also follow the conditions in section 3.
</p>
<p>You may also lend copies, under the same conditions stated above, and
you may publicly display copies.
</p>
</li><li> COPYING IN QUANTITY

<p>If you publish printed copies (or copies in media that commonly have
printed covers) of the Document, numbering more than 100, and the
Document&rsquo;s license notice requires Cover Texts, you must enclose the
copies in covers that carry, clearly and legibly, all these Cover
Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on
the back cover.  Both covers must also clearly and legibly identify
you as the publisher of these copies.  The front cover must present
the full title with all words of the title equally prominent and
visible.  You may add other material on the covers in addition.
Copying with changes limited to the covers, as long as they preserve
the title of the Document and satisfy these conditions, can be treated
as verbatim copying in other respects.
</p>
<p>If the required texts for either cover are too voluminous to fit
legibly, you should put the first ones listed (as many as fit
reasonably) on the actual cover, and continue the rest onto adjacent
pages.
</p>
<p>If you publish or distribute Opaque copies of the Document numbering
more than 100, you must either include a machine-readable Transparent
copy along with each Opaque copy, or state in or with each Opaque copy
a computer-network location from which the general network-using
public has access to download using public-standard network protocols
a complete Transparent copy of the Document, free of added material.
If you use the latter option, you must take reasonably prudent steps,
when you begin distribution of Opaque copies in quantity, to ensure
that this Transparent copy will remain thus accessible at the stated
location until at least one year after the last time you distribute an
Opaque copy (directly or through your agents or retailers) of that
edition to the public.
</p>
<p>It is requested, but not required, that you contact the authors of the
Document well before redistributing any large number of copies, to give
them a chance to provide you with an updated version of the Document.
</p>
</li><li> MODIFICATIONS

<p>You may copy and distribute a Modified Version of the Document under
the conditions of sections 2 and 3 above, provided that you release
the Modified Version under precisely this License, with the Modified
Version filling the role of the Document, thus licensing distribution
and modification of the Modified Version to whoever possesses a copy
of it.  In addition, you must do these things in the Modified Version:
</p>
<ol type="A" start="1">
<li> Use in the Title Page (and on the covers, if any) a title distinct
from that of the Document, and from those of previous versions
(which should, if there were any, be listed in the History section
of the Document).  You may use the same title as a previous version
if the original publisher of that version gives permission.

</li><li> List on the Title Page, as authors, one or more persons or entities
responsible for authorship of the modifications in the Modified
Version, together with at least five of the principal authors of the
Document (all of its principal authors, if it has fewer than five),
unless they release you from this requirement.

</li><li> State on the Title page the name of the publisher of the
Modified Version, as the publisher.

</li><li> Preserve all the copyright notices of the Document.

</li><li> Add an appropriate copyright notice for your modifications
adjacent to the other copyright notices.

</li><li> Include, immediately after the copyright notices, a license notice
giving the public permission to use the Modified Version under the
terms of this License, in the form shown in the Addendum below.

</li><li> Preserve in that license notice the full lists of Invariant Sections
and required Cover Texts given in the Document&rsquo;s license notice.

</li><li> Include an unaltered copy of this License.

</li><li> Preserve the section Entitled &ldquo;History&rdquo;, Preserve its Title, and add
to it an item stating at least the title, year, new authors, and
publisher of the Modified Version as given on the Title Page.  If
there is no section Entitled &ldquo;History&rdquo; in the Document, create one
stating the title, year, authors, and publisher of the Document as
given on its Title Page, then add an item describing the Modified
Version as stated in the previous sentence.

</li><li> Preserve the network location, if any, given in the Document for
public access to a Transparent copy of the Document, and likewise
the network locations given in the Document for previous versions
it was based on.  These may be placed in the &ldquo;History&rdquo; section.
You may omit a network location for a work that was published at
least four years before the Document itself, or if the original
publisher of the version it refers to gives permission.

</li><li> For any section Entitled &ldquo;Acknowledgements&rdquo; or &ldquo;Dedications&rdquo;, Preserve
the Title of the section, and preserve in the section all the
substance and tone of each of the contributor acknowledgements and/or
dedications given therein.

</li><li> Preserve all the Invariant Sections of the Document,
unaltered in their text and in their titles.  Section numbers
or the equivalent are not considered part of the section titles.

</li><li> Delete any section Entitled &ldquo;Endorsements&rdquo;.  Such a section
may not be included in the Modified Version.

</li><li> Do not retitle any existing section to be Entitled &ldquo;Endorsements&rdquo; or
to conflict in title with any Invariant Section.

</li><li> Preserve any Warranty Disclaimers.
</li></ol>

<p>If the Modified Version includes new front-matter sections or
appendices that qualify as Secondary Sections and contain no material
copied from the Document, you may at your option designate some or all
of these sections as invariant.  To do this, add their titles to the
list of Invariant Sections in the Modified Version&rsquo;s license notice.
These titles must be distinct from any other section titles.
</p>
<p>You may add a section Entitled &ldquo;Endorsements&rdquo;, provided it contains
nothing but endorsements of your Modified Version by various
parties&mdash;for example, statements of peer review or that the text has
been approved by an organization as the authoritative definition of a
standard.
</p>
<p>You may add a passage of up to five words as a Front-Cover Text, and a
passage of up to 25 words as a Back-Cover Text, to the end of the list
of Cover Texts in the Modified Version.  Only one passage of
Front-Cover Text and one of Back-Cover Text may be added by (or
through arrangements made by) any one entity.  If the Document already
includes a cover text for the same cover, previously added by you or
by arrangement made by the same entity you are acting on behalf of,
you may not add another; but you may replace the old one, on explicit
permission from the previous publisher that added the old one.
</p>
<p>The author(s) and publisher(s) of the Document do not by this License
give permission to use their names for publicity for or to assert or
imply endorsement of any Modified Version.
</p>
</li><li> COMBINING DOCUMENTS

<p>You may combine the Document with other documents released under this
License, under the terms defined in section 4 above for modified
versions, provided that you include in the combination all of the
Invariant Sections of all of the original documents, unmodified, and
list them all as Invariant Sections of your combined work in its
license notice, and that you preserve all their Warranty Disclaimers.
</p>
<p>The combined work need only contain one copy of this License, and
multiple identical Invariant Sections may be replaced with a single
copy.  If there are multiple Invariant Sections with the same name but
different contents, make the title of each such section unique by
adding at the end of it, in parentheses, the name of the original
author or publisher of that section if known, or else a unique number.
Make the same adjustment to the section titles in the list of
Invariant Sections in the license notice of the combined work.
</p>
<p>In the combination, you must combine any sections Entitled &ldquo;History&rdquo;
in the various original documents, forming one section Entitled
&ldquo;History&rdquo;; likewise combine any sections Entitled &ldquo;Acknowledgements&rdquo;,
and any sections Entitled &ldquo;Dedications&rdquo;.  You must delete all
sections Entitled &ldquo;Endorsements.&rdquo;
</p>
</li><li> COLLECTIONS OF DOCUMENTS

<p>You may make a collection consisting of the Document and other documents
released under this License, and replace the individual copies of this
License in the various documents with a single copy that is included in
the collection, provided that you follow the rules of this License for
verbatim copying of each of the documents in all other respects.
</p>
<p>You may extract a single document from such a collection, and distribute
it individually under this License, provided you insert a copy of this
License into the extracted document, and follow this License in all
other respects regarding verbatim copying of that document.
</p>
</li><li> AGGREGATION WITH INDEPENDENT WORKS

<p>A compilation of the Document or its derivatives with other separate
and independent documents or works, in or on a volume of a storage or
distribution medium, is called an &ldquo;aggregate&rdquo; if the copyright
resulting from the compilation is not used to limit the legal rights
of the compilation&rsquo;s users beyond what the individual works permit.
When the Document is included in an aggregate, this License does not
apply to the other works in the aggregate which are not themselves
derivative works of the Document.
</p>
<p>If the Cover Text requirement of section 3 is applicable to these
copies of the Document, then if the Document is less than one half of
the entire aggregate, the Document&rsquo;s Cover Texts may be placed on
covers that bracket the Document within the aggregate, or the
electronic equivalent of covers if the Document is in electronic form.
Otherwise they must appear on printed covers that bracket the whole
aggregate.
</p>
</li><li> TRANSLATION

<p>Translation is considered a kind of modification, so you may
distribute translations of the Document under the terms of section 4.
Replacing Invariant Sections with translations requires special
permission from their copyright holders, but you may include
translations of some or all Invariant Sections in addition to the
original versions of these Invariant Sections.  You may include a
translation of this License, and all the license notices in the
Document, and any Warranty Disclaimers, provided that you also include
the original English version of this License and the original versions
of those notices and disclaimers.  In case of a disagreement between
the translation and the original version of this License or a notice
or disclaimer, the original version will prevail.
</p>
<p>If a section in the Document is Entitled &ldquo;Acknowledgements&rdquo;,
&ldquo;Dedications&rdquo;, or &ldquo;History&rdquo;, the requirement (section 4) to Preserve
its Title (section 1) will typically require changing the actual
title.
</p>
</li><li> TERMINATION

<p>You may not copy, modify, sublicense, or distribute the Document
except as expressly provided under this License.  Any attempt
otherwise to copy, modify, sublicense, or distribute it is void, and
will automatically terminate your rights under this License.
</p>
<p>However, if you cease all violation of this License, then your license
from a particular copyright holder is reinstated (a) provisionally,
unless and until the copyright holder explicitly and finally
terminates your license, and (b) permanently, if the copyright holder
fails to notify you of the violation by some reasonable means prior to
60 days after the cessation.
</p>
<p>Moreover, your license from a particular copyright holder is
reinstated permanently if the copyright holder notifies you of the
violation by some reasonable means, this is the first time you have
received notice of violation of this License (for any work) from that
copyright holder, and you cure the violation prior to 30 days after
your receipt of the notice.
</p>
<p>Termination of your rights under this section does not terminate the
licenses of parties who have received copies or rights from you under
this License.  If your rights have been terminated and not permanently
reinstated, receipt of a copy of some or all of the same material does
not give you any rights to use it.
</p>
</li><li> FUTURE REVISIONS OF THIS LICENSE

<p>The Free Software Foundation may publish new, revised versions
of the GNU Free Documentation License from time to time.  Such new
versions will be similar in spirit to the present version, but may
differ in detail to address new problems or concerns.  See
<a href="http://www.gnu.org/copyleft/">http://www.gnu.org/copyleft/</a>.
</p>
<p>Each version of the License is given a distinguishing version number.
If the Document specifies that a particular numbered version of this
License &ldquo;or any later version&rdquo; applies to it, you have the option of
following the terms and conditions either of that specified version or
of any later version that has been published (not as a draft) by the
Free Software Foundation.  If the Document does not specify a version
number of this License, you may choose any version ever published (not
as a draft) by the Free Software Foundation.  If the Document
specifies that a proxy can decide which future versions of this
License can be used, that proxy&rsquo;s public statement of acceptance of a
version permanently authorizes you to choose that version for the
Document.
</p>
</li><li> RELICENSING

<p>&ldquo;Massive Multiauthor Collaboration Site&rdquo; (or &ldquo;MMC Site&rdquo;) means any
World Wide Web server that publishes copyrightable works and also
provides prominent facilities for anybody to edit those works.  A
public wiki that anybody can edit is an example of such a server.  A
&ldquo;Massive Multiauthor Collaboration&rdquo; (or &ldquo;MMC&rdquo;) contained in the
site means any set of copyrightable works thus published on the MMC
site.
</p>
<p>&ldquo;CC-BY-SA&rdquo; means the Creative Commons Attribution-Share Alike 3.0
license published by Creative Commons Corporation, a not-for-profit
corporation with a principal place of business in San Francisco,
California, as well as future copyleft versions of that license
published by that same organization.
</p>
<p>&ldquo;Incorporate&rdquo; means to publish or republish a Document, in whole or
in part, as part of another Document.
</p>
<p>An MMC is &ldquo;eligible for relicensing&rdquo; if it is licensed under this
License, and if all works that were first published under this License
somewhere other than this MMC, and subsequently incorporated in whole
or in part into the MMC, (1) had no cover texts or invariant sections,
and (2) were thus incorporated prior to November 1, 2008.
</p>
<p>The operator of an MMC Site may republish an MMC contained in the site
under CC-BY-SA on the same site at any time before August 1, 2009,
provided the MMC is eligible for relicensing.
</p>
</li></ol>

<span id="ADDENDUM_003a-How-to-use-this-License-for-your-documents"></span><h3 class="heading">ADDENDUM: How to use this License for your documents</h3>

<p>To use this License in a document you have written, include a copy of
the License in the document and put the following copyright and
license notices just after the title page:
</p>
<div class="example">
<pre class="example">  Copyright (C)  <var>year</var>  <var>your name</var>.
  Permission is granted to copy, distribute and/or modify this document
  under the terms of the GNU Free Documentation License, Version 1.3
  or any later version published by the Free Software Foundation;
  with no Invariant Sections, no Front-Cover Texts, and no Back-Cover
  Texts.  A copy of the license is included in the section entitled ``GNU
  Free Documentation License''.
</pre></div>

<p>If you have Invariant Sections, Front-Cover Texts and Back-Cover Texts,
replace the &ldquo;with&hellip;Texts.&rdquo; line with this:
</p>
<div class="example">
<pre class="example">    with the Invariant Sections being <var>list their titles</var>, with
    the Front-Cover Texts being <var>list</var>, and with the Back-Cover Texts
    being <var>list</var>.
</pre></div>

<p>If you have Invariant Sections without Cover Texts, or some other
combination of the three, merge those two alternatives to suit the
situation.
</p>
<p>If your document contains nontrivial examples of program code, we
recommend releasing these examples in parallel under your choice of
free software license, such as the GNU General Public License,
to permit their use in free software.
</p>


<hr>


</body>
</html>